Optimal stratification in outcome prediction using baseline information.
Ontology highlight
ABSTRACT:
SUBMITTER: Yong FH
PROVIDER: S-EPMC5793688 | biostudies-literature | 2016 Dec
REPOSITORIES: biostudies-literature
Publications
Optimal stratification in outcome prediction using baseline information.
Yong Florence H FH Tian Lu L Yu Sheng S Cai Tianxi T Wei L J LJ
Biometrika 20161208 4
A common practice in predictive medicine is to use current study data to construct a stratification procedure, which groups subjects according to baseline information and forms stratum-specific prevention or intervention strategies. A desirable stratification scheme would not only have small intra-stratum variation but also have a clinically meaningful discriminatory capability. We show how to obtain optimal stratification rules with such desirable properties from fitting a set of regression mod ...[more]
Similar Datasets
Project description:Responding to stimuli requires that organisms encode information about the external world. Not all parts of the input are important for behavior, and resource limitations demand that signals be compressed. Prediction of the future input is widely beneficial in many biological systems. We compute the trade-offs between representing the past faithfully and predicting the future using the information bottleneck approach, for input dynamics with different levels of complexity. For motion prediction, we show that, depending on the parameters in the input dynamics, velocity or position information is more useful for accurate prediction. We show which motion representations are easiest to re-use for accurate prediction in other motion contexts, and identify and quantify those with the highest transferability. For non-Markovian dynamics, we explore the role of long-term memory in shaping the internal representation. Lastly, we show that prediction in evolutionary population dynamics is linked to clustering allele frequencies into non-overlapping memories.
| S-EPMC7971903 | biostudies-literature
Project description:BACKGROUND:Prognostication is an essential tool for risk adjustment and decision making in the intensive care unit (ICU). Research into prognostication in ICU has so far been limited to data from admission or the first 24 hours. Most ICU admissions last longer than this, decisions are made throughout an admission, and some admissions are explicitly intended as time-limited prognostic trials. Despite this, temporal changes in prognostic ability during ICU admission has received little attention to date. Current predictive models, in the form of prognostic clinical tools, are typically derived from linear models and do not explicitly handle incremental information from trends. Machine learning (ML) allows predictive models to be developed which use non-linear predictors and complex interactions between variables, thus allowing incorporation of trends in measured variables over time; this has made it possible to investigate prognosis throughout an admission. METHODS AND FINDINGS:This study uses ML to assess the predictability of ICU mortality as a function of time. Logistic regression against physiological data alone outperformed APACHE-II and demonstrated several important interactions including between lactate & noradrenaline dose, between lactate & MAP, and between age & MAP consistent with the current sepsis definitions. ML models consistently outperformed logistic regression with Deep Learning giving the best results. Predictive power was maximal on the second day and was further improved by incorporating trend data. Using a limited range of physiological and demographic variables, the best machine learning model on the first day showed an area under the receiver-operator characteristic curve (AUC) of 0.883 (σ = 0.008), compared to 0.846 (σ = 0.010) for a logistic regression from the same predictors and 0.836 (σ = 0.007) for a logistic regression based on the APACHE-II score. Adding information gathered on the second day of admission improved the maximum AUC to 0.895 (σ = 0.008). Beyond the second day, predictive ability declined. CONCLUSION:This has implications for decision making in intensive care and provides a justification for time-limited trials of ICU therapy; the assessment of prognosis over more than one day may be a valuable strategy as new information on the second day helps to differentiate outcomes. New ML models based on trend data beyond the first day could greatly improve upon current risk stratification tools.
| S-EPMC6241126 | biostudies-other
Project description:Intradialytic hypotension (IDH) is a common side effect that occurs during hemodialysis and poses a great risk for dialysis patients. Many studies have been conducted so far to predict IDH, but most of these could not be applied in real-time because they used only underlying patient information or static patient disease information. In this study, we propose a multilayer perceptron (MP)-based IDH prediction model using heart rate (HR) information corresponding to time-series information and static data of patients. This study aimed to validate whether HR differences and HR slope information affect real-time IDH prediction in patients undergoing hemodialysis. Clinical data were collected from 80 hemodialysis patients from 9 September to 17 October 2020, in the artificial kidney room at Yeungnam University Medical Center (YUMC), Daegu, South Korea. The patients typically underwent hemodialysis 12 times during this period, 1 to 2 h per session. Therefore, the HR difference and HR slope information within up to 1 h before IDH occurrence were used as time-series input data for the MP model. Among the MP models using the number and data length of different hidden layers, the model using 60 min of data before the occurrence of two layers and IDH showed maximum performance, with an accuracy of 81.5%, a true positive rate of 73.8%, and positive predictive value of 87.3%. This study aimed to predict IDH in real-time by continuously supplying HR information to MP models along with static data such as age, diabetes, hypertension, and ultrafiltration. The current MP model was implemented using relatively limited parameters; however, its performance may be further improved by adding additional parameters in the future, further enabling real-time IDH prediction to play a supporting role for medical staff.
| S-EPMC9408052 | biostudies-literature
Project description:BackgroundNoninvasive risk stratification aims to detect abnormalities in the pathophysiological mechanisms underlying ventricular arrhythmias. We studied the predictive value of repeating risk stratification in patients with an implantable cardioverter-defibrillator (ICD).MethodsThe EUTrigTreat clinical study was a prospective multicenter trial including ischemic and nonischemic cardiomyopathies and arrhythmogenic heart disease. Left ventricular ejection fraction ≤40% (LVEF), premature ventricular complexes >400/24 hr (PVC), non-negative microvolt T-wave alternans (MTWA), and abnormal heart rate turbulence (HRT) were considered high risk. Tests were repeated within 12 months after inclusion. Adjusted Cox regression analysis was performed for mortality and appropriate ICD shocks.ResultsIn total, 635 patients had analyzable baseline data with a median follow-up of 4.4 years. Worsening of LVEF was associated with increased mortality (HR 3.59, 95% CI 1.17-11.04), as was consistent abnormal HRT (HR 8.34, 95%CI 1.06-65.54). HRT improvement was associated with improved survival when compared to consistent abnormal HRT (HR 0.10, 95%CI 0.01-0.82). For appropriate ICD shocks, a non-negative MTWA test or high PVC count at any moment was associated with increased arrhythmic risk independent of the evolution of test results (worsening: HR 3.76 (95%CI 1.43-9.88) and HR 2.50 (95%CI 1.15-5.46); improvement: HR 2.80 (95%CI 1.03-7.61) and HR 2.45 (95%CI 1.07-5.62); consistent: HR 2.47 (95%CI 0.95-6.45) and HR 2.40 (95%CI 1.33-4.33), respectively). LVEF improvement was associated with a lower arrhythmic risk (HR 0.34, 95%CI 0.12-0.94).ConclusionsRepeating LVEF and HRT improved the prediction of mortality, whereas stratification of ventricular arrhythmias may be improved by repeating LVEF measurements, MTWA and ECG Holter monitoring.
| S-EPMC7679829 | biostudies-literature
Project description:Melanoma is currently divided on a genetic level according to mutational status. However, this classification does not optimally predict prognosis. In prior studies, we have defined gene expression phenotypes (high-immune, pigmentation, proliferative and normal-like), which are predictive of survival outcome as well as informative of biology. Herein, we employed a population-based metastatic melanoma cohort and external cohorts to determine the prognostic and predictive significance of the gene expression phenotypes. We performed expression profiling on 214 cutaneous melanoma tumors and found an increased risk of developing distant metastases in the pigmentation (HR, 1.9; 95% CI, 1.05-3.28; P=0.03) and proliferative (HR, 2.8; 95% CI, 1.43-5.57; P=0.003) groups as compared to the high-immune response group. Further genetic characterization of melanomas using targeted deep-sequencing revealed similar mutational patterns across these phenotypes. We also used publicly available expression profiling data from melanoma patients treated with targeted or vaccine therapy in order to determine if our signatures predicted therapeutic response. In patients receiving targeted therapy, melanomas resistant to targeted therapy were enriched in the MITF-low proliferative subtype as compared to pre-treatment biopsies (P=0.02). In summary, the melanoma gene expression phenotypes are highly predictive of survival outcome and can further help to discriminate patients responding to targeted therapy.
| S-EPMC4494939 | biostudies-literature
Project description:ObjectiveIdentifying subgroups of low back pain (LBP) has the potential to improve prediction of clinical outcomes. Risk stratification is one such strategy that identifies similar characteristics indicative of a common clinical outcome trajectory. The purpose of this study was to determine if an empirically derived subgrouping approach based on physical impairment measures improves information provided from the STarT Back Tool (SBT).MethodsAt baseline in this secondary analysis of a cohort study, patients (N = 144) receiving physical therapy for LBP completed the SBT and tests (active lumbar flexion, extension, lateral bending, and passive straight-leg raise) from a validated physical impairment index. Clinical outcomes were assessed at 4 weeks and included the Numerical Pain Rating Scale and Oswestry Disability Index. Exploratory hierarchical agglomerative cluster analysis identified empirically derived subgroups based on physical impairment measures. Independent samples t testing and chi-square analysis were used to assess baseline subgroup differences in demographic and clinical measures. Spearman rho correlation coefficient was used to assess baseline SBT risk and impairment subgroup relationships, and a 3-way mixed-model ANOVA was used to assessed SBT risk and impairment subgroup relationships with clinical outcomes at 4 weeks.ResultsTwo physical impairment-based subgroups emerged from cluster analysis: (1) low-risk impairment (n = 119, 81.5%), characterized by greater lumbar mobility; and (2) high-risk impairment (n = 25, 17.1%), characterized by less lumbar mobility. A weak, positive relationship was observed between baseline SBT risk and impairment subgroups (rs = .170). An impairment-by-SBT risk-by-time interaction effect was observed for Oswestry Disability Index scores but not for Numerical Pain Rating Scale scores at 4 weeks.ConclusionsPhysical impairment subgroups were not redundant with SBT risk categories and could improve prediction of 4-week LBP disability outcomes. Physical impairment subgroups did not improve the prediction of 4-week pain intensity scores.ImpactSubgroups based on physical impairment and psychosocial risk could lead to better prediction of LBP disability outcomes and eventually allow for treatment options tailored to physical and psychosocial risk.
| S-EPMC8179624 | biostudies-literature
Project description:In the status forecasting problem, classification models such as logistic regression with input variables such as physiological, diagnostic, and treatment variables are typical ways of modeling. However, the parameter value and model performance differ among individuals with different baseline information. To cope with these difficulties, a subgroup analysis is conducted, in which models' ANOVA and rpart are proposed to explore the influence of baseline information on the parameters and model performance. The results show that the logistic regression model achieves satisfactory performance, which is generally higher than 0.95 in AUC and around 0.9 in F1 and balanced accuracy. The subgroup analysis presents the prior parameter values for monitoring variables including SpO2, milrinone, non-opioid analgesics and dobutamine. The proposed method can be used to explore variables that are and are not medically related to the baseline variables.
| S-EPMC9601351 | biostudies-literature
Project description:Readily available proxies for the time of disease onset such as the time of the first diagnostic code can lead to substantial risk prediction error if performing analyses based on poor proxies. Due to the lack of detailed documentation and labor intensiveness of manual annotation, it is often only feasible to ascertain for a small subset the current status of the disease by a follow-up time rather than the exact time. In this paper, we aim to develop risk prediction models for the onset time efficiently leveraging both a small number of labels on the current status and a large number of unlabeled observations on imperfect proxies. Under a semiparametric transformation model for onset and a highly flexible measurement error model for proxy onset time, we propose the semisupervised risk prediction method by combining information from proxies and limited labels efficiently. From an initially estimator solely based on the labeled subset, we perform a one-step correction with the full data augmenting against a mean zero rank correlation score derived from the proxies. We establish the consistency and asymptotic normality of the proposed semisupervised estimator and provide a resampling procedure for interval estimation. Simulation studies demonstrate that the proposed estimator performs well in a finite sample. We illustrate the proposed estimator by developing a genetic risk prediction model for obesity using data from Mass General Brigham Healthcare Biobank.
| S-EPMC9741856 | biostudies-literature
Project description:Current treatment of chronic hepatitis C virus (HCV) infection has limited efficacy -especially among genotype 1 infected patients-, is costly, and involves severe side effects. Thus, predicting non-response is of major interest for both patient wellbeing and health care expense. At present, treatment cannot be individualized on the basis of any baseline predictor of response. We aimed to identify pre-treatment clinical and virological parameters associated with treatment failure, as well as to assess whether therapy outcome could be predicted at baseline.Forty-three HCV subtype 1b (HCV-1b) chronically infected patients treated with pegylated-interferon alpha plus ribavirin were retrospectively studied (21 responders and 22 non-responders). Host (gender, age, weight, transaminase levels, fibrosis stage, and source of infection) and viral-related factors (viral load, and genetic variability in the E1-E2 and Core regions) were assessed. Logistic regression and discriminant analyses were used to develop predictive models. A "leave-one-out" cross-validation method was used to assess the reliability of the discriminant models.Lower alanine transaminase levels (ALT, p=0.009), a higher number of quasispecies variants in the E1-E2 region (number of haplotypes, nHap_E1-E2) (p=0.003), and the absence of both amino acid arginine at position 70 and leucine at position 91 in the Core region (p=0.039) were significantly associated with treatment failure. Therapy outcome was most accurately predicted by discriminant analysis (90.5% sensitivity and 95.5% specificity, 85.7% sensitivity and 81.8% specificity after cross-validation); the most significant variables included in the predictive model were the Core amino acid pattern, the nHap_E1-E2, and gamma-glutamyl transferase and ALT levels.Discriminant analysis has been shown as a useful tool to predict treatment outcome using baseline HCV genetic variability and host characteristics. The discriminant models obtained in this study led to accurate predictions in our population of Spanish HCV-1b treatment naïve patients.
| S-EPMC2994723 | biostudies-literature
Project description:BACKGROUND: Predicting the function of a protein is one of the major challenges in the post-genomic era where a large number of protein sequences of unknown function are accumulating rapidly. Lectins are the proteins that specifically recognize and bind to carbohydrate moieties present on either proteins or lipids. Cancerlectins are those lectins that play various important roles in tumor cell differentiation and metastasis. Although the two types of proteins are linked, still there is no computational method available that can distinguish cancerlectins from the large pool of non-cancerlectins. Hence, it is imperative to develop a method that can distinguish between cancer and non-cancerlectins. RESULTS: All the models developed in this study are based on a non-redundant dataset containing 178 cancerlectins and 226 non-cancerlectins in which no two sequences have more than 50% sequence similarity. We have applied the similarity search based technique, i.e. BLAST, and achieved a maximum accuracy of 43.25%. The amino acids compositional analysis have shown that certain residues (e.g. Leucine, Proline) were preferred in cancerlectins whereas some other (e.g. Asparatic acid, Asparagine) were preferred in non-cancerlectins. It has been found that the PROSITE domain "Crystalline beta gamma" was abundant in cancerlectins whereas domains like "SUEL-type lectin domain" were found mainly in non-cancerlectins. An SVM-based model has been developed to differentiate between the cancer and non-cancerlectins which achieved a maximum Matthew's correlation coefficient (MCC) value of 0.32 with an accuracy of 64.84%, using amino acid compositions. We have developed a model based on dipeptide compositions which achieved an MCC value of 0.30 with an accuracy of 64.84%. Thereafter, we have developed models based on split compositions (2 and 4 parts) and achieved an MCC value of 0.31, 0.32 with accuracies of 65.10% and 66.09%, respectively. An SVM model based on Position Specific Scoring Matrix (PSSM), generated by PSI-BLAST, was developed and achieved an MCC value of 0.36 with an accuracy of 68.34%. Finally, we have integrated the PROSITE domain information with PSSM and developed an SVM model that has achieved an MCC value of 0.38 with 69.09% accuracy. CONCLUSION: BLAST has been found inefficient to distinguish between cancer and non-cancerlectins. We analyzed the protein sequences of cancer and non-cancerlectins and identified interesting patterns. We have been able to identify PROSITE domains that are preferred in cancer and non-cancerlectins and thus provided interesting insights into the two types of proteins. The method developed in this study will be useful for researchers studying cancerlectins, lectins and cancer biology. The web-server based on the above study, is available at http://www.imtech.res.in/raghava/cancer_pred/
| S-EPMC3161874 | biostudies-other