Dataset Information

A systematic review of the clinical application of data-driven population segmentation analysis.

ABSTRACT:

Background

Data-driven population segmentation analysis utilizes data analytics to divide a heterogeneous population into parsimonious and relatively homogenous groups with similar healthcare characteristics. It is a promising patient-centric analysis that enables effective integrated healthcare interventions specific for each segment. Although widely applied, there is no systematic review on the clinical application of data-driven population segmentation analysis.

Methods

We carried out a systematic literature search using PubMed, Embase and Web of Science following PRISMA criteria. We included English peer-reviewed articles that applied data-driven population segmentation analysis on empirical health data. We summarized the clinical settings in which segmentation analysis was applied, compared and contrasted strengths, limitations, and practical considerations of different segmentation methods, and assessed the segmentation outcome of all included studies. The studies were assessed by two independent reviewers.

Results

We retrieved 14,514 articles and included 216 articles. Data-driven population segmentation analysis was widely used in different clinical contexts. 163 studies examined the general population while 53 focused on specific population with certain diseases or conditions, including psychological, oncological, respiratory, cardiovascular, and gastrointestinal conditions. Variables used for segmentation in the studies are heterogeneous. Most studies (n = 170) utilized secondary data in community settings (n = 185). The most common segmentation method was latent class/profile/transition/growth analysis (n = 96) followed by K-means cluster analysis (n = 60) and hierarchical analysis (n = 50), each having its advantages, disadvantages, and practical considerations. We also identified key criteria to evaluate a segmentation framework: internal validity, external validity, identifiability/interpretability, substantiality, stability, actionability/accessibility, and parsimony.

Conclusions

Data-driven population segmentation has been widely applied and holds great potential in managing population health. The evaluations of segmentation outcome require the interplay of data analytics and subject matter expertise. The optimal framework for segmentation requires further research.

SUBMITTER: Yan S

PROVIDER: S-EPMC6215625 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundHealthcare needs-based population segmentation is a promising approach for enabling the development and evaluation of integrated healthcare service models that meet healthcare needs. However, healthcare policymakers interested in understanding adult population healthcare needs may not be aware of suitable population segmentation tools available for use in the literature and barring better-known alternatives, may reinvent the wheel by creating and validating their own tools rather than adapting available tools in the literature. Therefore, we undertook a systematic review to identify all available tools which operationalize healthcare need-based population segmentation, to help inform policymakers developing population-level health service programmes.MethodsUsing search terms reflecting concepts of population, healthcare need and segmentation, we systematically reviewed and included articles containing healthcare need-based adult population segmentation tools in PubMed, CINAHL and Web of Science databases. We included tools comprising mutually exclusive segments with prognostic value for clinically relevant outcomes. An updated secondary search on the PubMed database was also conducted as the last search was conducted 2?years ago. All identified tools were characterized in terms of segment formulation, segmentation base, whether they received peer-reviewed validation, requirement for comprehensive electronic medical records, proprietary status and number of segments.ResultsA total of 16 unique tools were identified from systematically reviewing 9970 articles. Peer-reviewed validation studies were found for 9 of these tools.Discussion and conclusionsThe underlying segmentation basis of most identified tools was found to be conceptually comparable to each other which suggests a broad recognition of archetypical patient overall healthcare need profiles. While many tools operate based on administrative record data, it is noted that healthcare systems without comprehensive electronic medical records would benefit from tools which segment populations through primary data collection. Future work could therefore include development and validation of such primary data collection-based tools. While this study is limited by exclusion of non-English literature, the identified and characterized tools will nonetheless facilitate efforts by policymakers to improve patient-centred care through development and evaluation of services tailored for specific populations segmented by these tools.

Project description:IntroductionPulse harmonic analysis is a quantitative and objective methodology within traditional Chinese medicine (TCM) used to evaluate pulse characteristics. However, interpreting pulse wave data is challenging due to its inherent complexity. This study aims to provide a comprehensive review and comparison of existing human pulse wave harmonic analysis methods to elucidate their patterns and characteristics.MethodsA systematic review of clinical research reports published from 1990 to 2021 was conducted, focusing on variations in harmonic characteristics across different medical conditions and physiological states. Keyword searches included terms related to analysis methods (e.g., "Pulse Spectrum," "harmonic analysis," "harmonic index") and measured indicators (e.g., "vascular response," "PPG," "Photoplethysmography," "aortic," "arterial," "blood pressure"). Supplementary research using PubMed's Mesh terms specifically targeted "Pulse wave analysis" within the methods and statistical analysis domain. Articles were filtered based on predefined criteria, including human participants and research related to pulse pressure or vascular volume changes. Conference papers, animal studies, and irrelevant research were excluded, with literature evaluation scales selected based on the retrieved research reports.ResultsInitially, 6487 research reports were identified, and after screening, 50 reports were included in the review. The analysis revealed that low-frequency harmonics increase following vigorous activity or sympathetic excitation but decrease during rest or parasympathetic excitation. Cardiovascular patients exhibited elevated first harmonics associated with the liver meridian, while diabetes patients displayed weakened third harmonics related to the spleen meridian. Liver dysfunction was linked to changes in the first harmonic, and cancer patients showed signs of liver and kidney yin deficiency in the first and second harmonics. These findings underscore the potential of harmonic analysis for TCM disease diagnosis and organ assessment. Moreover, individuals with conditions such as liver dysfunction, cancer, and gynecological disorders displayed distinct intensity patterns across harmonics one through ten compared to healthy controls, albeit with some variations. Heterogeneity in these studies mainly stemmed from differences in measurement methods and study populations. Additionally, research suggested that factors like blood circulation and cognitive activity influenced harmonic intensity.ConclusionsIn summary, this report consolidates prior research on pulse wave harmonics analysis, revealing unique patterns associated with various physiological conditions. Despite limitations, such as limited sample sizes in previous studies, the observed associations between physiological states and harmonics hold promise for potential clinical applications. This study lays a solid foundation for future applications of arterial wave harmonics analysis, promoting wider adoption of this analytical approach.

Project description:BackgroundAccording to current classification systems, patients with major depressive disorder (MDD) may have very different combinations of symptoms. This symptomatic diversity hinders the progress of research into the causal mechanisms and treatment allocation. Theoretically founded subtypes of depression such as atypical, psychotic, and melancholic depression have limited clinical applicability. Data-driven analyses of symptom dimensions or subtypes of depression are scarce. In this systematic review, we examine the evidence for the existence of data-driven symptomatic subtypes of depression.MethodsWe undertook a systematic literature search of MEDLINE, PsycINFO and Embase in May 2012. We included studies analyzing the depression criteria of the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) of adults with MDD in latent variable analyses.ResultsIn total, 1176 articles were retrieved, of which 20 satisfied the inclusion criteria. These reports described a total of 34 latent variable analyses: 6 confirmatory factor analyses, 6 exploratory factor analyses, 12 principal component analyses, and 10 latent class analyses. The latent class techniques distinguished 2 to 5 classes, which mainly reflected subgroups with different overall severity: 62 of 71 significant differences on symptom level were congruent with a latent class solution reflecting severity. The latent class techniques did not consistently identify specific symptom clusters. Latent factor techniques mostly found a factor explaining the variance in the symptoms depressed mood and interest loss (11 of 13 analyses), often complemented by psychomotor retardation or fatigue (8 of 11 analyses). However, differences in found factors and classes were substantial.ConclusionsThe studies performed to date do not provide conclusive evidence for the existence of depressive symptom dimensions or symptomatic subtypes. The wide diversity of identified factors and classes might result either from the absence of patterns to be found, or from the theoretical and modeling choices preceding analysis.

Project description:Background: Latent class analysis (LCA) can be used to identify subgroups of children with similar patterns of mental health symptoms and/or strengths. The method is becoming more commonly used in child mental health research, but there are reservations about the replicability, reliability, and validity of findings. Objective: A systematic literature review was conducted to investigate the extent to which LCA has been used to study population mental health in children, and whether replicable, reliable and valid findings have been demonstrated. Methods: Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. A search of literature, published between January 1998 and December 2017, was carried out using MEDLINE, EMBASE, PsycInfo, Scopus, ERIC, ASSIA, and Google Scholar. A total of 2,748 studies were initially identified, of which 23 were eligible for review. The review examined the methods which studies had used to choose the number of mental health classes, the classes that they found, and whether there was evidence for the validity and reliability of the classes. Results: Reviewed studies used LCA to investigate both disparate mental health symptoms, and those associated with specific disorders. The corpus of studies using similar indicators was small. Differences in the criteria used to select the final LCA model were found between studies. All studies found meaningful or useful subgroups, but there were differences in the extent to which the validity and reliability of classes were explicitly demonstrated. Conclusions : LCA is a useful tool for studying and classifying child mental health at the population level. Recommendations are made to improve the application and reporting of LCA and to increase confidence in findings in the future, including use of a range of indices and criteria when enumerating classes, clear reporting of methods for replicability, and making efforts to establish the validity and reliability of identified classes.

Project description:BACKGROUND:Clinical narratives represent the main form of communication within health care, providing a personalized account of patient history and assessments, and offering rich information for clinical decision making. Natural language processing (NLP) has repeatedly demonstrated its feasibility to unlock evidence buried in clinical narratives. Machine learning can facilitate rapid development of NLP tools by leveraging large amounts of text data. OBJECTIVE:The main aim of this study was to provide systematic evidence on the properties of text data used to train machine learning approaches to clinical NLP. We also investigated the types of NLP tasks that have been supported by machine learning and how they can be applied in clinical practice. METHODS:Our methodology was based on the guidelines for performing systematic reviews. In August 2018, we used PubMed, a multifaceted interface, to perform a literature search against MEDLINE. We identified 110 relevant studies and extracted information about text data used to support machine learning, NLP tasks supported, and their clinical applications. The data properties considered included their size, provenance, collection methods, annotation, and any relevant statistics. RESULTS:The majority of datasets used to train machine learning models included only hundreds or thousands of documents. Only 10 studies used tens of thousands of documents, with a handful of studies utilizing more. Relatively small datasets were utilized for training even when much larger datasets were available. The main reason for such poor data utilization is the annotation bottleneck faced by supervised machine learning algorithms. Active learning was explored to iteratively sample a subset of data for manual annotation as a strategy for minimizing the annotation effort while maximizing the predictive performance of the model. Supervised learning was successfully used where clinical codes integrated with free-text notes into electronic health records were utilized as class labels. Similarly, distant supervision was used to utilize an existing knowledge base to automatically annotate raw text. Where manual annotation was unavoidable, crowdsourcing was explored, but it remains unsuitable because of the sensitive nature of data considered. Besides the small volume, training data were typically sourced from a small number of institutions, thus offering no hard evidence about the transferability of machine learning models. The majority of studies focused on text classification. Most commonly, the classification results were used to support phenotyping, prognosis, care improvement, resource management, and surveillance. CONCLUSIONS:We identified the data annotation bottleneck as one of the key obstacles to machine learning approaches in clinical NLP. Active learning and distant supervision were explored as a way of saving the annotation efforts. Future research in this field would benefit from alternatives such as data augmentation and transfer learning, or unsupervised learning, which do not require data annotation.

Project description:BackgroundData-driven gating (DDG) can improve PET quantitation and alleviate many issues with patient motion. However, misregistration between DDG-PET and CT may occur due to the distinct temporal resolutions of PET and CT and can be mitigated by DDG-CT. Here, the effects of misregistration and respiratory motion on PET quantitation and lesion segmentation were assessed with a new DDG-PET/CT method.MethodsA low-dose cine-CT was acquired in misregistered regions to enable both average CT (ACT) and DDG-CT. The following were compared: (1) baseline PET/CT, (2) PET/ACT (attenuation correction, AC = ACT), (3) DDG-PET (AC = helical CT), and (4) DDG-PET/CT (AC = DDG-CT). For DDG-PET, end-expiration (EE) data were derived from 50% of the total PET data at 30% from end-inspiration. For DDG-CT, EE phase CT data were extracted from cine-CT data by lung Hounsfield unit (HU) value and body contour. A total of 91 lesions from 16 consecutive patients were assessed for changes in standard uptake value (SUV), lesion glycolysis (LG), lesion volume, centroid-to-centroid distance (CCD), and DICE coefficients.ResultsRelative to baseline PET/CT, median changes in SUVmax ± σ for all 91 lesions were 20 ± 43%, 26 ± 23%, and 66 ± 66%, respectively, for PET/ACT, DDG-PET, and DDG-PET/CT. Median changes in lesion volume were 0 ± 58%, - 36 ± 26%, and - 26 ± 40%. LG for individual lesions increased for PET/ACT and decreased for DDG-PET, but was not different for DDG-PET/CT. Changes in mean HU from baseline PET/CT were dramatic for most lesions in both PET/ACT and DDG-PET/CT, especially for lesions with mean HU < 0 at baseline. CCD and DICE were both affected more by motion correction with DDG-PET than improved registration with ACT or DDG-CT.ConclusionAs misregistration becomes more prominent, the impact of motion correction with DDG-PET is diminished. The potential benefits of DDG-PET toward accurate lesion segmentation and quantitation could only be fully realized when combined with DDG-CT. These results impress upon the necessity of ensuring both misregistration and motion correction are accounted for together to optimize the clinical utility of PET/CT.

Project description:ImportancePain is a silent global epidemic impacting approximately a third of the population. Pharmacological and surgical interventions are primary modes of treatment. Cognitive/behavioural management approaches and interventional pain management strategies are approaches that have been used to assist with the management of chronic pain. Accurate data collection and reporting treatment outcomes are vital to addressing the challenges faced. In light of this, we conducted a systematic evaluation of the current digital application landscape within chronic pain medicine.ObjectiveThe primary objective was to consider the prevalence of digital application usage for chronic pain management. These digital applications included mobile apps, web apps, and chatbots.Data sourcesWe conducted searches on PubMed and ScienceDirect for studies that were published between 1st January 1990 and 1st January 2021.Study selectionOur review included studies that involved the use of digital applications for chronic pain conditions. There were no restrictions on the country in which the study was conducted. Only studies that were peer-reviewed and published in English were included. Four reviewers had assessed the eligibility of each study against the inclusion/exclusion criteria. Out of the 84 studies that were initially identified, 38 were included in the systematic review.Data extraction and synthesisThe AMSTAR guidelines were used to assess data quality. This assessment was carried out by 3 reviewers. The data were pooled using a random-effects model.Main outcomes and measuresBefore data collection began, the primary outcome was to report on the standard mean difference of digital application usage for chronic pain conditions. We also recorded the type of digital application studied (e.g., mobile application, web application) and, where the data was available, the standard mean difference of pain intensity, pain inferences, depression, anxiety, and fatigue.Results38 studies were included in the systematic review and 22 studies were included in the meta-analysis. The digital interventions were categorised to web and mobile applications and chatbots, with pooled standard mean difference of 0.22 (95% CI: -0.16, 0.60), 0.30 (95% CI: 0.00, 0.60) and -0.02 (95% CI: -0.47, 0.42) respectively. Pooled standard mean differences for symptomatologies of pain intensity, depression, and anxiety symptoms were 0.25 (95% CI: 0.03, 0.46), 0.30 (95% CI: 0.17, 0.43) and 0.37 (95% CI: 0.05, 0.69), respectively. A sub-group analysis was conducted on pain intensity due to the heterogeneity of the results (I 2 = 82.86%; p = 0.02). After stratifying by country, we found that digital applications were more likely to be effective in some countries (e.g., United States, China) than others (e.g., Ireland, Norway).Conclusions and relevanceThe use of digital applications in improving pain-related symptoms shows promise, but further clinical studies would be needed to develop more robust applications.Systematic review registrationhttps://www.crd.york.ac.uk/prospero/, identifier: CRD42021228343.