Machine Learning Predicts the Fall Risk of Total Hip Arthroplasty Patients Based on Wearable Sensor Instrumented Performance Tests

BackgroundThe prevalence of falls affects the wellbeing of aging adults and places an economic burden on the healthcare system. Integration of wearable sensors into existing fall risk assessment tools enables objective data collection that describes the functional ability of patients. In this study, supervised machine learning was applied to sensor-derived metrics to predict the fall risk of patients following total hip arthroplasty.MethodsAt preoperative, 2-week, and 6-week postoperative appointments, patients (n = 72) were instrumented with sensors while they performed the timed-up-and-go walking test. Preoperative and 2-week postoperative data were used to form the feature sets and 6-week total times were used as labels. Support vector machine and linear discriminant analysis classifier models were developed and tested on various combinations of feature sets and feature reduction schemes. Using a 10-fold leave-some-subjects-out testing scheme, the accuracy, sensitivity, specificity, and area under the receiver-operator curve (AUC) were evaluated for all models.ResultsA high performance model (accuracy = 0.87, sensitivity = 0.97, specificity = 0.46, AUC = 0.82) was obtained with a support vector machine classifier using sensor-derived metrics from only the preoperative appointment. An overall improved performance (accuracy = 0.90, sensitivity = 0.93, specificity = 0.59, AUC = 0.88) was achieved with a linear discriminant analysis classifier when 2-week postoperative data were added to the preoperative data.ConclusionThe high accuracy of the fall risk prediction models is valuable for patients, clinicians, and the healthcare system. High-risk patients can implement preventative measures and low-risk patients can be directed to enhanced recovery care programs.     

Clinical prediction models to predict the risk of multiple binary outcomes: a comparison of approaches

Clinical prediction models (CPMs) can predict clinically relevant outcomes or events. Typically, prognostic CPMs are derived to predict the risk of a single future outcome. However, there are many medical applications where two or more outcomes are of interest, meaning this should be more widely reflected in CPMs so they can accurately estimate the joint risk of multiple outcomes simultaneously. A potentially naïve approach to multi‐outcome risk prediction is to derive a CPM for each outcome separately, then multiply the predicted risks. This approach is only valid if the outcomes are conditionally independent given the covariates, and it fails to exploit the potential relationships between the outcomes. This paper outlines several approaches that could be used to develop CPMs for multiple binary outcomes. We consider four methods, ranging in complexity and conditional independence assumptions: namely, probabilistic classifier chain, multinomial logistic regression, multivariate logistic regression, and a Bayesian probit model. These are compared with methods that rely on conditional independence: separate univariate CPMs and stacked regression. Employing a simulation study and real‐world example, we illustrate that CPMs for joint risk prediction of multiple outcomes should only be derived using methods that model the residual correlation between outcomes. In such a situation, our results suggest that probabilistic classification chains, multinomial logistic regression or the Bayesian probit model are all appropriate choices. We call into question the development of CPMs for each outcome in isolation when multiple correlated or structurally related outcomes are of interest and recommend more multivariate approaches to risk prediction.     

Preliminary observations on the application of the multiple inocula (replicator) method for carbon substrate utilization studies of Alcaligenes faecalis.

The use of the replicator technique in evaluating carbon source utilization by 55 cultures of Alcaligenes faecalis was examined. Six of the 20 substrates tested were utilized by the bacterial cultures. In this study, the reproducibility of the technique was 100%.     

An evaluation of three methods for determining colloid osmotic pressure

Plasma colloid osmotic pressure (COP) is an important determinant in edema formation. Three methods for assessing the COP were evaluated. Direct measurement of COP using the 4420 Wescor Colloid Osmometer was compared to the estimation of COP from both serum total protein and total serum solids (TSS) determinations. Blood samples from twenty adult patients (mean age = 64 years) undergoing cardiopulmonary bypass surgery were collected for COP assessment. Sample collection was performed prior to heparinization/hemodilution, during hypothermic bypass and at the conclusion of bypass following protamine administration. The results obtained from each method were analyzed by a two-way analysis of variance. The Bonferroni technique was used for comparison of sample means when the difference was significant (p less than 0.05). Correlations were reported by linear regression analysis. A statistically significant difference (p less than 0.01) was found between the three methods. A regression equation for the estimation of COP from total serum solids is offered: COP = (3.02 * TSS) + 0.65. Prospective clinical testing between the direct COP measurement and the estimation of COP from TSS using the equation (n = 38) revealed a significant correlation (R2 = .932) and no significant difference between the two (p greater than 0.05).     

Comparison of gaseous microemboli counts in arterial, simultaneous and venous heat exchange with a hollow fiber membrane oxygenator

Potential sources of gaseous microemboli during cardiopulmonary bypass are varied. However, it is known that membrane oxygenators generate fewer gaseous microemboli than bubble oxygenators and that bubblers cannot utilize arterial heat exchange without generating significant gaseous microemboli during rewarming. A membrane oxygenator utilizing simultaneous gas and heat exchange raises the concern that concurrent gas and heat exchange would result in a higher production of gaseous microemboli compared to conventional venous heat exchange devices. This in vitro study compared venous, simultaneous, arterial and control (venous) heat exchanger gaseous microemboli counts during rewarming. No significant difference was found between the four heat exchangers when comparing inlet and outlet gaseous microemboli counts. This in vitro study suggests that there is no difference in gaseous microemboli generation when varying the position of the heat exchanger in the extracorporeal circuit incorporating a microporous membrane oxygenator.     

Life-long absence of sexual drive in a woman associated with 5-dihydrotestosterone deficiency

A case of a woman who presented with life-long absence of sexual drive is reported. Although her history contained psychological factors that might have been etiologic in her presentation, she had not responded to sex therapy or cognitive-behavior psychotherapy undertaken at other clinics. When seen by the author, she was found to have poorly developed external genitalia. The only abnormality discovered on endocrinological evaluation was a reduced level of 5 dihydrotestosterone. Treatment with dihydrotestosterone gel applied to her vulva generated sexual drive and her ability to become sexually aroused.     

Prostate cancer treatment and ten-year survival among group/staff HMO and fee-for-service Medicare patients.

OBJECTIVE: To compare treatment patterns and the ten-year survival of prostate cancer patients in two large, nonprofit, group/staff HMOs to those of patients receiving care in the fee-for-service health setting. DATA SOURCES/STUDY DESIGN: A cohort of men age 65 and over diagnosed with prostate cancer between 1985 and the end of 1992 and followed through 1994. Subjects (n = 21,741) were ascertained by two population-based tumor registries covering the greater San Francisco-Oakland and Seattle-Puget Sound areas. Linkage of registry data with Medicare claims data and with HMO inpatient utilization data allowed the determination of health plan enrollment and the measurement of comorbid conditions. Multivariate regression models were used to examine HMO versus FFS treatment and survival differences adjusting for sociodemographic and clinical characteristics. PRINCIPAL FINDINGS: Among cases with non-metastatic prostate cancer, HMO patients were more likely than FFS patients to receive aggressive therapy (either prostatectomy or radiation) in San Francisco-Oakland (odds ratio [OR] = 1.69, 95% CI = 1.46-1.96) but not in Seattle (OR = 1.15, 0.93-1.43). Among men receiving aggressive therapy, HMO cases were three to five times more likely to receive radiation therapy than prostatectomy. Overall mortality was equivalent over ten years (HMO versus FFS mortality risk ratio [RR] = 1.01, 0.94-1.08), but prostate cancer mortality was higher for HMO cases than for FFS cases (RR = 1.25, 1.13-1.39). CONCLUSION: Despite marked treatment differences for clinically localized prostate cancer, overall ten-year survival for patients enrolled in two nonprofit group/staff HMOs was equivalent to survival among patients receiving care in the FFS setting, even after adjustment for sociodemographic and clinical characteristics. Similar overall but better prostate cancer-specific survival among FFS patients is most plausibly explained by differences between the HMO and FFS patients in both tumor characteristics and unmeasured patient selection factors.     

Commercialization of health services: implications for the laboratories

The commercialization of health services has wide ranging implications for all medical specialties as well as for patients. Factors that must be considered include not only the financial implications, but also questions of quality and academic interests such as teaching and training. Laboratories must provide a service that the purchaser wishes to buy and must be successful in overcoming competition from the private sector. Each component part of the overall service must be analyzed in order that the laboratory is efficiently structured to provide an optimum service. A good understanding of management issues and a flexible approach are paramount in the provision of efficient, cost-effective and quality service for the ultimate benefit of the patient.