Aravindharamanan S, profile picture
Uploaded byAravindharamanan S
128 views

Sun==big data analytics for health care

The document provides an overview of big data analytics for healthcare. It begins with motivating examples that demonstrate how big data can help improve healthcare outcomes and lower costs. It then discusses the main sources of healthcare data, including structured EHR data like billing codes, labs, and medications, as well as unstructured clinical notes. The document outlines challenges in analyzing these different types of complex healthcare data. It also introduces a healthcare analytics platform that can extract and select features from various data sources to build predictive models. Finally, it discusses techniques for clinical text mining, including named entity recognition and negation analysis.

Download to read offline
1 / 112
2 / 112
3 / 112
4 / 112
5 / 112
6 / 112
7 / 112
8 / 112
9 / 112
10 / 112
11 / 112
12 / 112
13 / 112
14 / 112
15 / 112
16 / 112
17 / 112
18 / 112
19 / 112
20 / 112
21 / 112
22 / 112
23 / 112
24 / 112
25 / 112
26 / 112
27 / 112
28 / 112
29 / 112
30 / 112
31 / 112
32 / 112
33 / 112
34 / 112
35 / 112
36 / 112
37 / 112
38 / 112
39 / 112
40 / 112
41 / 112
42 / 112
43 / 112
44 / 112
45 / 112
46 / 112
47 / 112
48 / 112
49 / 112
50 / 112
51 / 112
52 / 112
53 / 112
54 / 112
55 / 112
56 / 112
57 / 112
58 / 112
59 / 112
60 / 112
61 / 112
62 / 112
63 / 112
64 / 112
65 / 112
66 / 112
67 / 112
68 / 112
69 / 112
70 / 112
71 / 112
72 / 112
73 / 112
74 / 112
75 / 112
76 / 112
77 / 112
78 / 112
79 / 112
80 / 112
81 / 112
82 / 112
83 / 112
84 / 112
85 / 112
86 / 112
87 / 112
88 / 112
89 / 112
90 / 112
91 / 112
92 / 112
93 / 112
94 / 112
95 / 112
96 / 112
97 / 112
98 / 112
99 / 112
100 / 112
101 / 112
102 / 112
103 / 112
104 / 112
105 / 112
106 / 112
107 / 112
108 / 112
109 / 112
110 / 112
111 / 112
112 / 112
1 Big Data Analytics for Healthcare Chandan K. Reddy Department of Computer Science Wayne State University Tutorial presentation at the SIAM International Conference on Data Mining, Austin, TX, 2013. The updated tutorial slides are available at http://dmkd.cs.wayne.edu/TUTORIAL/Healthcare/ Jimeng Sun Healthcare Analytics Department IBM TJ Watson Research Center
2 Motivation Hersh, W., Jacko, J. A., Greenes, R., Tan, J., Janies, D., Embi, P. J., & Payne, P. R. (2011). Health-care hit or miss? Nature, 470(7334), 327. In 2012, worldwide digital healthcare data was estimated to be equal to 500 petabytes and is expected to reach 25,000 petabytes in 2020. Can we learn from the past to become better in the future ?? Healthcare Data is becoming more complex !!
3 Organization of this Tutorial  Introduction  Motivating Examples  Sources and Techniques for Big Data in Healthcare – Structured EHR Data – Unstructured Clinical Notes – Medical Imaging Data – Genetic Data – Other Data (Epidemiology & Behavioral)  Final Thoughts and Conclusion
4 INTRODUCTION
5 Definition of Big Data  A collection of large and complex data sets which are difficult to process using common database management tools or traditional data processing applications. The challenges include capturing, storing, searching, sharing & analyzing. The four dimensions (V’s) of Big Data Big data is not just about size. • Finds insights from complex, noisy, heterogeneous, longitudinal, and voluminous data. • It aims to answer questions that were previously unanswered. BIG DATA Velocity Veracity Variety Volume  “Big data refers to the tools, processes and procedures allowing an organization to create, manipulate, and manage very large data sets and storage facilities” – according to zdnet.com
6 Reasons for Growing Complexity/Abundance of Healthcare Data  Standard medical practice is moving from relatively ad-hoc and subjective decision making to evidence-based healthcare.  More incentives to professionals/hospitals to use EHR technology. Additional Data Sources  Development of new technologies such as capturing devices, sensors, and mobile applications.  Collection of genomic information became cheaper.  Patient social communications in digital forms are increasing.  More medical knowledge/discoveries are being accumulated.
7 Big Data Challenges in Healthcare  Inferring knowledge from complex heterogeneous patient sources. Leveraging the patient/data correlations in longitudinal records.  Understanding unstructured clinical notes in the right context.  Efficiently handling large volumes of medical imaging data and extracting potentially useful information and biomarkers.  Analyzing genomic data is a computationally intensive task and combining with standard clinical data adds additional layers of complexity.  Capturing the patient’s behavioral data through several sensors; their various social interactions and communications.
8  Take advantage of the massive amounts of data and provide right intervention to the right patient at the right time.  Personalized care to the patient.  Potentially benefit all the components of a healthcare system i.e., provider, payer, patient, and management. Electronic Health Records Evidence + Insights Improved outcomes through smarter decisions Lower costs Big Data Analytics Overall Goals of Big Data Analytics in Healthcare BehavioralGenomic Public Health
9 Purpose of this Tutorial Two-fold objectives:  Introduce the data mining researchers to the sources available and the possible challenges and techniques associated with using big data in healthcare domain.  Introduce Healthcare analysts and practitioners to the advancements in the computing field to effectively handle and make inferences from voluminous and heterogeneous healthcare data. The ultimate goal is to bridge data mining and medical informatics communities to foster interdisciplinary works between the two communities. PS: Due to the broad nature of the topic, the primary emphasis will be on introducing healthcare data repositories, challenges, and concepts to data scientists. Not much focus will be on describing the details of any particular techniques and/or solutions.
10 Disclaimers  Being a recent and growing topic, there might be several other resources that might not be covered here.  Presentation here is more biased towards the data scientists’ perspective and may be less towards the healthcare management or healthcare provider’s perspective.  Some of the website links provided might become obsolete in the future. This tutorial is prepared in early 2013.  Since this topic contains a wide varieties of problems, there might be some aspects of healthcare that might not be covered in the tutorial.
11 MOTIVATING EXAMPLES
12 EXAMPLE 1: Heritage Health Prize  Over $30 billion was spent on unnecessary hospital admissions. Goals:  Identify patients at high-risk and ensure they get the treatment they need.  Develop algorithms to predict the number of days a patient will spend in a hospital in the next year. Outcomes:  Health care providers can develop new strategies to care for patients before its too late reduces the number of unnecessary hospitalizations.  Improving the health of patients while decreasing the costs of care.  Winning solutions use a combination of several predictive models. http://www.heritagehealthprize.com
13 EXAMPLE 2: Penalties for Poor Care - 30-Day Readmissions  Hospitalizations account for more than 30% of the 2 trillion annual cost of healthcare in the United States. Around 20% of all hospital admissions occur within 30 days of a previous discharge. – not only expensive but are also potentially harmful, and most importantly, they are often preventable.  Medicare penalizes hospitals that have high rates of readmissions among patients with heart failure, heart attack, and pneumonia.  Identifying patients at risk of readmission can guide efficient resource utilization and can potentially save millions of healthcare dollars each year.  Effectively making predictions from such complex hospitalization data will require the development of novel advanced analytical models.
14 EXAMPE 3: White House unveils BRAIN Initiative  The US President unveiled a new bold $100 million research initiative designed to revolutionize our understanding of the human brain. BRAIN (Brain Research through Advancing Innovative Neurotechnologies) Initiative.  Find new ways to treat, cure, and even prevent brain disorders, such as Alzheimer’s disease, epilepsy, and traumatic brain injury.  “Every dollar we invested to map the human genome returned $140 to our economy... Today, our scientists are mapping the human brain to unlock the answers to Alzheimer’s.” -- President Barack Obama, 2013 State of the Union.  “advances in "Big Data" that are necessary to analyze the huge amounts of information that will be generated; and increased understanding of how thoughts, emotions, actions and memories are represented in the brain.” : NSF  Joint effort by NSF, NIH, DARPA, and other private partners. http://www.whitehouse.gov/infographics/brain-initiative
15 EXAMPLE 4: GE Head Health Challenge Challenge 1: Methods for Diagnosis and Prognosis of Mild Traumatic Brain Injuries. Challenge 2: The Mechanics of Injury: Innovative Approaches For Preventing And Identifying Brain Injuries. In Challenge 1, GE and the NFL will award up to $10M for two types of solutions: Algorithms and Analytical Tools, and Biomarkers and other technologies. A total of $60M in funding over a period of 4 years.
16 Healthcare Continuum Sarkar, Indra Neil. "Biomedical informatics and translational medicine." Journal of Translational Medicine 8.1 (2010): 22.
17 Data Collection and Analysis Jensen, Peter B., Lars J. Jensen, and Søren Brunak. "Mining electronic health records: towards better research applications and clinical care." Nature Reviews Genetics (2012). Effectively integrating and efficiently analyzing various forms of healthcare data over a period of time can answer many of the impending healthcare problems.
18 Organization of this Tutorial  Introduction  Motivating Examples  Sources and Techniques for Big Data in Healthcare – Structured EHR Data – Unstructured Clinical Notes – Medical Imaging Data – Genetic Data – Other Data (Epidemiology & Behavioral)  Final Thoughts and Conclusion
19 SOURCES AND TECHNIQUES FOR BIG DATA IN HEALTHCARE
20 Outline Electronic Health Records (EHR) data Healthcare Analytic Platform Resources
21 ELECTRONIC HEALTH RECORDS (EHR) DATA
22 Data Health data Genomic data • DNA sequences Clinical data • Structured EHR • Unstructured EHR • Medical Images Behavior data • Social network data • Mobility sensor data
23 Billing data - ICD codes  ICD stands for International Classification of Diseases  ICD is a hierarchical terminology of diseases, signs, symptoms, and procedure codes maintained by the World Health Organization (WHO)  In US, most people use ICD-9, and the rest of world use ICD-10  Pros: Universally available  Cons: medium recall and medium precision for characterizing patients • (250) Diabetes mellitus • (250.0) Diabetes mellitus without mention of complication • (250.1) Diabetes with ketoacidosis • (250.2) Diabetes with hyperosmolarity • (250.3) Diabetes with other coma • (250.4) Diabetes with renal manifestations • (250.5) Diabetes with ophthalmic manifestations • (250.6) Diabetes with neurological manifestations • (250.7) Diabetes with peripheral circulatory disorders • (250.8) Diabetes with other specified manifestations • (250.9) Diabetes with unspecified complication
24 Billing data – CPT codes  CPT stands for Current Procedural Terminology created by the American Medical Association  CPT is used for billing purposes for clinical services  Pros: High precision  Cons: Low recall Codes for Evaluation and Management: 99201-99499 (99201 - 99215) office/other outpatient services (99217 - 99220) hospital observation services (99221 - 99239) hospital inpatient services (99241 - 99255) consultations (99281 - 99288) emergency dept services (99291 - 99292) critical care services …
25 Lab results  The standard code for lab is Logical Observation Identifiers Names and Codes (LOINC®)  Challenges for lab – Many lab systems still use local dictionaries to encode labs – Diverse numeric scales on different labs • Often need to map to normal, low or high ranges in order to be useful for analytics – Missing data • not all patients have all labs • The order of a lab test can be predictive, for example, BNP indicates high likelihood of heart failure Time Lab Value 1996-03-15 12:50:00.0 CO2 29.0 1996-03-15 12:50:00.0 BUN 16.0 1996-03-15 12:50:00.0 HDL-C 37.0 1996-03-15 12:50:00.0 K 4.5 1996-03-15 12:50:00.0 Cl 102.0 1996-03-15 12:50:00.0 Gluc 86.0
26 Medication  Standard code is National Drug Code (NDC) by Food and Drug Administration (FDA), which gives a unique identifier for each drug – Not used universally by EHR systems – Too specific, drugs with the same ingredients but different brands have different NDC  RxNorm: a normalized naming system for generic and branded drugs by National Library of Medicine  Medication data can vary in EHR systems – can be in both structured or unstructured forms  Availability and completeness of medication data vary – Inpatient medication data are complete, but outpatient medication data are not – Medication usually only store prescriptions but we are not sure whether patients actually filled those prescriptions
27 Clinical notes  Clinical notes contain rich and diverse source of information  Challenges for handling clinical notes – Ungrammatical, short phrases – Abbreviations – Misspellings – Semi-structured information • Copy-paste from other structure source – Lab results, vital signs • Structured template: – SOAP notes: Subjective, Objective, Assessment, Plan
28 Summary of common EHR data ICD CPT Lab Medication Clinical notes Availability High High High Medium Medium Recall Medium Poor Medium Inpatient: High Outpatient: Variable Medium Precision Medium High High Inpatient: High Outpatient: Variable Medium high Format Structured Structured Mostly structured Structured and unstructured Unstructured Pros Easy to work with, a good approximation of disease status Easy to work with, high precision High data validity High data validity More details about doctors’ thoughts Cons Disease code often used for screening, therefore disease might not be there Missing data Data normalization and ranges Prescribed not necessary taken Difficult to process Joshua C. Denny Chapter 13: Mining Electronic Health Records in the Genomics Era. PLoS Comput Biol. 2012 December; 8(12):
29 Analytic Platform Large-scale Healthcare Analytic Platform
30 Analytic Platform Healthcare Analytics Information Extraction Feature Selection Predictive Modeling
31 Analytic Platform Healthcare Analytics Information Extraction Feature Selection Predictive Modeling Structured EHR Structured EHR Feature extraction Feature extraction Unstructured EHR Unstructured EHR Patient Representation Patient Representation Feature Selection Feature Selection ContextContext ClassificationClassification RegressionRegression Patient Similarity Patient Similarity
32 Analytic Platform Healthcare Analytics Information Extraction Feature Selection Predictive Modeling Structured EHR Structured EHR Feature extraction Feature extraction Unstructured EHR Unstructured EHR Patient Representation Patient Representation Feature Selection Feature Selection ContextContext ClassificationClassification RegressionRegression Patient Similarity Patient Similarity
33 CLINICAL TEXT MINING
34 Text Mining in Healthcare  Text mining – Information Extraction • Name Entity Recognition – Information Retrieval  Clinical text vs. Biomedical text – Biomedical text: medical literatures (well-written medical text) – Clinical text is written by clinicians in the clinical settings • Meystre et al. Extracting Information from Textual Documents in the Electronic Health Record: A Review of Recent Research. IMIA 2008 • Zweigenbaum et al. Frontiers of biomedical text mining: current progress, BRIEFINGS IN BIOINFORMATICS. VOL 8. NO 5. 358-375 • Cohen and Hersh, A survey of current work in biomedical text mining. BRIEFINGS IN BIOINFORMATICS. VOL 6. NO 1. 57–71.
35 Auto-Coding: Extracting Codes from Clinical Text  Problem – Automatically assign diagnosis codes to clinical text  Significance – The cost is approximately $25 billion per year in the US  Available Data – Medical NLP Challenges from 2007 • Subsections from radiology reports: clinical history and impression  Potential Evaluation Metric: – F-measure = 2P*R/(P+R), where P is precision, and R is recall.  Example References • Aronson et al. From indexing the biomedical literature to coding clinical text: experience with MTI and machine learning approaches. BioNLP 2007 • Crammer et al. Automatic Code Assignment to Medical Text. BioNLP 2007 • Friedman C, Shagina L, Lussier Y, Hripcsak G. Automated encoding of clinical documents based on natural language processing. JAMIA. 2004:392-402 • Pakhomov SV, Buntrock JD, Chute CG. Automating the assignment of diagnosis codes to patient encounters using example-based and machine learning techniques. JAMIA 2006:516-25.
36 Context Analysis - Negation  Negation: e.g., ...denies chest pain… – NegExpander [1] achieves 93% precision on mammographic reports – NegEx [2] uses regular expression and achieves 94.5% specificity and 77.8% sensitivity – NegFinder [3] uses UMLS and regular expression, and achieves 97.7 specificity and 95.3% sensitivity when analyzing surgical notes and discharge summaries – A hybrid approach [4] uses regular expression and grammatical parsing and achieves 92.6% sensitivity and 99.8% specificity 1. Aronow DB, Fangfang F, Croft WB. Ad hoc classification of radiology reports. JAMIA 1999:393-411 2. Chapman et al. A simple algorithm for identifying negated findings and diseases in discharge summaries. JBI 2001:301-10. 3. Mutalik PG, et al. Use of general-purpose negation detection to augment concept indexing of medical documents: a quantitative study using the UMLS. JAMIA 2001:598-609. 4. Huang Y, Lowe HJ. A novel hybrid approach to automated negation detection in clinical radiology reports. JAMIA 2007
37 Context Analysis - Temporality  Temporality: e.g., …fracture of the tibia 2 years ago – TimeText [1] can detect temporal relations with 93.2% recall and 96.9% precision on 14 discharge summaries – Context [2] is an extension of NegEx, which identifies • negations (negated, affirmed), • temporality (historical, recent, hypothetical) • experiencer (patient, other) 1. Zhou et al. The Evaluation of a Temporal Reasoning System in Processing Clinical Discharge Summaries. JAMIA 2007. 2. Chapman W, Chu D, Dowling JN. ConText: An Algorithm for Identifying Contextual Features from Clinical Text. BioNLP 2007 Trigger TerminationClinical concepts Scope of the context v v v v
38 CASE 1: CASE BASED RETRIEVAL Sondhi P, Sun J, Zhai C, Sorrentino R, Kohn MS. Leveraging medical thesauri and physician feedback for improving medical literature retrieval for case queries. JAMIA. 2012
39 Input: Case Query Patient with smoking habit and weight loss. The frontal and lateral chest X rays show a mass in the posterior segment of the right upper lobe as well as a right hilar enlargement and obliteration of the right paratracheal stripe. On the chest CT the contours of the mass are lobulated with heterogeneous enhancement.Enlarged mediastinal and hilar lymph nodes are present.
40 Goal: Find Relevant Research Articles to a Query Additional Related Information Disease MeSH
41 Challenge 1: Query Weighing Patient with smoking habit and weight loss. The frontal and lateral chest X rays show a mass in the posterior segment of the right upper lobe as well as a right hilar enlargement and obliteration of the right paratracheal stripe. On the chest CT the contours of the mass are lobulated with heterogeneous enhancement. Enlarged mediastinal and hilar lymph nodes are present.  Queries are long  Not all words useful  IDF does not reflect importance  Semantics decide weight
42 Method: Semantic Query Weighing Identify important UMLS Semantic Types based on their definition Assign higher weights to query words under these types Disease or syndrome, Body part organ or organ component, Sign or symptom, Finding, Acquired abnormality, Congenital abnormality, Mental or behavioral dysfunction, Neoplasm, Pharmacologic substance, Individual Behavior Included UMLS semantic types
43  Matching variants  “x ray”, “x-rays”, “x rays”  Matching synonyms  “CT” or “x rays”  Knowledge gap Challenge 2: Vocabulary Gap Patient with smoking habit and weight loss. The frontal and lateral chest X rays show a mass in the posterior segment of the right upper lobe as well as a right hilar enlargement and obliteration of the right paratracheal stripe. On the chest CT the contours of the mass are lobulated with heterogeneous enhancement.Enlarged mediastinal and hilar lymph nodes are present.
44 Method: Additional Query Keywords  Asked physicians to provide additional keywords  Adding them with low weight helps  Any potential diagnosis keywords help greatly  Gives us insights into better query formulation Female patient, 25 years old, with fatigue and a swallowing disorder (dysphagia worsening during a meal). The frontal chest X-ray shows opacity with clear contours in contact with the right heart border. Right hilar structures are visible through the mass. The lateral X-ray confirms the presence of a mass in the anterior mediastinum. On CT images, the mass has a relatively homogeneous tissue density. Additional keywords: Thymoma, Lymphoma, Dysphagia, Esophageal obstruction, Myasthenia gravis, Fatiguability, Ptosis
45 Challenge 3: Pseudo-Feedback General vocabulary gap solution: – Apply Pseudo-Relevance Feedback What if very few of the top N are relevant? No idea which keywords to pick up
46 Method: Medical Subject Heading (MeSH) Feedback Any case related query usually relates only to handful of conditions How to guess the condition of the query? –Select MeSH terms from top N=10 ranked documents –Select MeSH terms covering most query keywords –Use them for feedback
47 Method: MeSH Feedback Lung NeoplasmsDoc 1 BronchitisDoc 2 Cystic FibrosisDoc 3 Lung NeoplasmsDoc 4 HepatitisDoc 5
48 Method 2: MeSH Feedback Lung NeoplasmsDoc 1 BronchitisDoc 2 Cystic FibrosisDoc 3 Lung NeoplasmsDoc 4 HepatitisDoc 5 Filtration List Lung Neoplasms Bronchitis
49 Method 2: MeSH Feedback Lung NeoplasmsDoc 1 BronchitisDoc 2 Cystic FibrosisDoc 3 Lung NeoplasmsDoc 4 HepatitisDoc 5 Filtration List Lung Neoplasms Bronchitis Reduce Weight Leave Unchanged Reduce Weight
50 Method: MeSH Feedback Lung NeoplasmsDoc 1 BronchitisDoc 2 Cystic FibrosisDoc 3 Lung NeoplasmsDoc 4 HepatitisDoc 5 Filtration List Lung Neoplasms Bronchitis Reduce Weight Leave Unchanged Reduce Weight Cystic FibrosisDoc 3 Lung NeoplasmsDoc 4 HepatitisDoc 5 Lung NeoplasmsDoc 1 BronchitisDoc 2
51 Retrieval Results Best performing run Data + Knowledge helps!
52 CASE 2: HEART FAILURE SIGNS AND SYMPTOMS Roy J. Byrd, Steven R. Steinhubl, Jimeng Sun, Shahram Ebadollahi, Walter F. Stewart. Automatic identification of heart failure diagnostic criteria, using text analysis of clinical notes from electronic health records. International Journal of Medical Informatics 2013
53 Framingham HF Signs and Symptoms  Framingham criteria for HF* are signs and symptoms that are documented even at primary care visits * McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: the Framingham study. N Engl J Med. 1971;285(26):1441-6.
54 Natural Language Processing (NLP) Pipeline  Criteria extraction comes from sentence level.  Encounter label comes from the entire note. Roy J. Byrd, Steven R. Steinhubl, Jimeng Sun, Shahram Ebadollahi, Walter F. Stewart. Automatic identification of heart failure diagnostic criteria, using text analysis of clinical notes from electronic health records. International Journal of Medical Informatics 2013
55 Performance on Encounter Level on Test Set  Machine learning method: decision tree  Rule-based method is to construct grammars by computational linguists  Manually constructed rules are more accurate but more effort to construct than automatic rules from learning a decision tree 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Recall Precision F-Score Recall Precision F-Score Machine-learning method Rule-based method Overall Affir m e d Denied
56 Potential Impact on Evidence-based Therapies 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% 70.00% Preceding Framingham diagnosis After Framingham diagnosis After clinical diagnosis  Applying text mining to extract Framingham symptoms can help trigger early intervention 3,168 patients eventually all diagnosed with HF Vhavakrishnan R, Steinhubl SR, Sun J, et al. Potential impact of predictive models for early detection of heart failure on the initiation of evidence-based therapies. J Am Coll Cardiol. 2012;59(13s1):E949-E949. No symptoms Framingham symptoms Clinical diagnosis Opportunity for early intervention
57 Analytic Platform Healthcare Analytics Information Extraction Feature Selection Predictive Modeling Structured EHR Structured EHR Feature extraction Feature extraction Unstructured EHR Unstructured EHR Patient Representation Patient Representation Feature Selection Feature Selection ContextContext ClassificationClassification RegressionRegression Patient Similarity Patient Similarity
58 KNOWLEDGE+DATA FEATURE SELECTION Jimeng Sun, Jianying Hu, Dijun Luo, Marianthi Markatou, Fei Wang, Shahram Ebadollahi, Steven E. Steinhubl, Zahra Daar, Walter F. Stewart. Combining Knowledge and Data Driven Insights for Identifying Risk Factors using Electronic Health Records. AMIA (2012).
59 Combining Knowledge- and Data-driven Risk Factors Knowledge Data Combination Risk factor augmentation Risk factor augmentation Combined risk factors Combined risk factors Knowledge base Knowledge base Risk factor gathering Risk factor gathering Knowledge risk factors Knowledge risk factors Clinical dataClinical data Data processing Data processing Potential risk factors Potential risk factors Target condition Target condition Risk factor augmentation Risk factor augmentation
60 Risk Factor Augmentation  Model Accuracy: – The selected risk factors are highly predictive of the target condition – Sparse feature selection through L1 regularization  Minimal Correlations: – Between data driven risk factors and knowledge driven risk factors – Among the data driven risk factors Model error Correlation among data-driven features Correlation between data- and knowledge- driven features Sparse Penalty Dijun Luo, Fei Wang, Jimeng Sun, Marianthi Markatou, Jianying Hu,Shahram Ebadollahi, SOR: Scalable Orthogonal Regression for Low-Redundancy Feature Selection and its Healthcare Applications. SDM’12
61 Prediction Results using Selected Features  AUC significantly improves as complementary data driven risk factors are added into existing knowledge based risk factors.  A significant AUC increase occurs when we add first 50 data driven features +200+150 +100 +50 all knowledge features+diabetes +Hypertension CAD 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0 100 200 300 400 500 600 AUC Number of features Jimeng Sun, Jianying Hu, Dijun Luo, Marianthi Markatou, Fei Wang, Shahram Ebadollahi, Steven E. Steinhubl, Zahra Daar, Walter F. Stewart. Combining Knowledge and Data Driven Insights for Identifying Risk Factors using Electronic Health Records. AMIA (2012).
62 Clinical Validation of Data-driven Features  9 out of 10 are considered relevant to HF  The data driven features are complementary to the existing knowledge-driven features
63 Analytic Platform Healthcare Analytics Information Extraction Feature Selection Predictive Modeling Structured EHR Structured EHR Feature extraction Feature extraction Unstructured EHR Unstructured EHR Patient Representation Patient Representation Feature Selection Feature Selection ContextContext ClassificationClassification RegressionRegression Patient Similarity Patient Similarity
64 PREDICTIVE MODEL
65 Anatomy of Clinical Predictive Model  Prediction Models –Continuous outcome: Regression –Categorical outcome: Classification • Logistic regression –Survival outcome • Cox Proportional Hazard Regression –Patient Similarity  Case study: Heart failure onset prediction
66 PATIENT SIMILARITY
67 Intuition of Patient Similarity Patient Doctor Similarity search
68 Intuition of Patient Similarity Patient Doctor
69 Summary on Patient Similarity  Patient similarity learns a customized distance metric for a specific clinical context  Extension 1: Composite distance integration (Comdi) [SDM’11a] – How to jointly learn a distance by multiple parties without data sharing?  Extension 2: Interactive metric update (iMet) [SDM’11b] – How to interactively update an existing distance measure? 1. Jimeng Sun, Fei Wang, Jianying Hu, Shahram Edabollahi: Supervised patient similarity measure of heterogeneous patient records. SIGKDD Explorations 14(1): 16-24 (2012) 2. Fei Wang, Jimeng Sun, Shahram Ebadollahi: Integrating Distance Metrics Learned from Multiple Experts and its Application in Inter-Patient Similarity Assessment. SDM 2011: 59-70 56 3. Fei Wang, Jimeng Sun, Jianying Hu, Shahram Ebadollahi: iMet: Interactive Metric Learning in Healthcare Applications. SDM 2011: 944-955
70 CASE STUDY: HEART FAILURE PREDICTION
71 Motivations for Early Detection of Heart Failure  Heart failure (HF) is a complex disease  Huge Societal Burden  For payers – Reduce cost and hospitalization – Improve the existing clinical guidance of HF prevention  For providers – Slow or potentially reverse disease progress – Improve quality of life, reduce mortality
72 Predictive Modeling Study Design  Goal: Classify HF cases against control patients  Population – 50,625 Patients (Geisinger Clinic PCPs) – Cases: 4,644 case patients – Controls 45,981 matched on age, gender and clinic Cases Controls
73 Predictive Modeling Setup  We define – Diagnosis date and index date – Prediction and observation windows  Features are constructed from the observation window and predict HF onset after the prediction window Index date Diagnosis date Observation Window Prediction Window
74 Features  We construct over 20K features of different types  Through feature selection and generalization, we result in the following predictive features Feature type Cardinality Predictive Features DIAGNOSIS 17,322 Diabetes, CHD, hypertensions, valvular disease, left ventricular hypertrophy, angina, atrial fibrillation, MI, COPD Demographics 11 Age, race, gender, smoking status Framingham 15 rales, cardiomegaly, acute pulmonary edema, HJReflex, ankle edema, nocturnal cough, DOExertion, hepatomegaly, pleural effusion Lab 1,264 eGFR, LVEF, albumin, glucose, cholesterol, creatinine, cardiomegaly, heart rate, hemoglobin Medication 3,922 antihypertensive, lipid-lowering, CCB, ACEI, ARB, beta blocker, diuretic, digitalis, antiarrhythmic Vital 6 blood pressure and heart rate
75 Prediction Performance on Different Prediction Windows  Setting: observation window = 720 days, classifiers = random forest, evaluation mechanism = 10-fold cross-validation for 10 times  Observation:  AUC slowly decreases as the prediction window increases 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 0 90 180 270 360 450 540 630 720 AUC Prediction Window (days) HF Onset (ObservationWindow=720, PredictionWindow=variable)
76 Prediction Performance on Different Observation Windows  Setting: prediction window= 180 days, classifiers= random forest, evaluation mechanism =10-fold cross-validation  Observation:  AUC increases as the observation window increases. i.e., more data for a longer period of time will lead to better performance of the predictive model  Combined features performed the best at observation window = 720 days 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 30 90 180 270 360 450 540 630 720 810 900 AUC Observation Window (days) HF Onset (ObservationWindow=variable, PredictionWindow=180)
77 Analytic Platform Healthcare Analytics Information Extraction Feature Selection Predictive Modeling Structured EHR Structured EHR Feature extraction Feature extraction Unstructured EHR Unstructured EHR Patient Representation Patient Representation Feature Selection Feature Selection ContextContext ClassificationClassification RegressionRegression Patient Similarity Patient Similarity
78 RESOURCES
79 Unstructured Clinical Data Dataset Link Description i2b2 Informatics for Integrating Biology & the Bedside https://www.i2b2.org/NLP/Dat aSets/Main.php Clinical notes used for clinical NLP challenges • 2006 Deidentification and Smoking Challenge • 2008 Obesity Challenge • 2009 Medication Challenge • 2010 Relations Challenge • 2011 Co-reference Challenge Computational Medicine center http://computationalmedicine. org/challenge/previous Classifying Clinical Free Text Using Natural Language Processing
80 Structured EHR Dataset Link Description Texas Hospital Inpatient Discharge http://www.dshs.state.tx.us/thcic/ho spitals/Inpatientpudf.shtm Patient: hospital location, admission type/source, claims, admit day, age, icd9 codes + surgical codes Framingham Health Care Data Set http://www.framinghamheartstudy.o rg/share/index.html Genetic dataset for cardiovascular disease Medicare Basic Stand Alone Claim Public Use Files http://resdac.advantagelabs.com/c ms-data/files/bsa-puf Inpatient, skilled nursing facility, outpatient, home health agency, hospice, carrier, durable medical equipment, prescription drug event, and chronic conditions on an aggregate level VHA Medical SAS Datasets http://www.virec.research.va.gov/M edSAS/Overview.htm Patient care encounters primarily for Veterans: inpatient/outpatient data from VHA facilities Nationwide Inpatient Sample http://www.hcup- us.ahrq.gov/nisoverview.jsp Discharge data from 1051 hospitals in 45 states with diagnosis, procedures, status, demographics, cost, length of stay CA Patient Discharge Data http://www.oshpd.ca.gov/HID/Produ cts/PatDischargeData/PublicDataS et/index.html Discharge data for licensed general acute hospital in CA with demographic, diagnostic and treatment information, disposition, total charges MIMIC II Clinical Database http://mimic.physionet.org/database .html ICU data including demographics, diagnosis, clinical measurements, lab results, interventions, notes Thanks to Prof. Joydeep Ghosh from UT Austin for providing this information
81 Software  MetaMap maps biomedical text to UMLS metathesaurus – Developed by NLM for parsing medical article not clinical notes – http://metamap.nlm.nih.gov/  cTAKES: clinical Text Analysis and Knowledge Extraction System – Using Unstructured Information Management Architecture (UIMA) framework and OpenNLP toolkit – http://ctakes.apache.org/
82 Organization of this Tutorial  Introduction  Motivating Examples  Sources and Techniques for Big Data in Healthcare – Structured EHR Data – Unstructured Clinical Notes – Medical Imaging Data – Genetic Data – Other Data (Epidemiology & Behavioral)  Final Thoughts and Conclusion
83 MEDICAL IMAGE DATA
84 Image Data is Big !!!  By 2015, the average hospital will have two-thirds of a petabyte (665 terabytes) of patient data, 80% of which will be unstructured image data like CT scans and X-rays.  Medical Imaging archives are increasing by 20%-40%  PACS (Picture Archival & Communication Systems) system is used for storage and retrieval of the images. Image Source: http://medcitynews.com/2013/03/the-body-in-bytes-medical-images-as-a-source-of-healthcare-big-data-infographic/
85 Popular Imaging Modalities in Healthcare Domain Computed Tomography (CT)  The main challenge with the image data is that it is not only huge, but is also high-dimensional and complex.  Extraction of the important and relevant features is a daunting task.  Many research works applied image features to extract the most relevant images for a given query. Positron Emission Tomography (PET) Magnetic Resonance Imaging (MRI) Image Source: Wikipedia
86 Medical Image Retrieval System Biomedical Image Database Testing Phase Feature Extraction Algorithms for learning or similarity computations Final Trained Models Query Image Retrieval System Query Results Performance Evaluation (Precision-Recall) Recall Precision Training Phase
87 Content-based Image Retrieval  Two components – Image features/descriptors - bridging the gap between the visual content and its numerical representation. – These representations are designed to encode color and texture properties of the image, the spatial layout of objects, and various geometric shape characteristics of perceptually coherent structures. – Assessment of similarities between image features based on mathematical analyses, which compare descriptors across different images. – Vector affinity measures such as Euclidean distance, Mahalanobis distance, KL divergence, Earth Mover’s distance are amongst the widely used ones.
88 Medical Image Features Photo-metric features exploit color and texture cues and they are derived directly from raw pixel intensities. Geometric features: cues such as edges, contours, joints, polylines, and polygonal regions. Akgül, Ceyhun Burak, et al. "Content-based image retrieval in radiology: current status and future directions." Journal of Digital Imaging 24.2 (2011): 208-222. Müller, Henning, et al. "A review of content-based image retrieval systems in medical applications-clinical benefits and future directions." International journal of medical informatics 73.1 (2004): 1-24. • A suitable shape representation should be extracted from the pixel intensity information by region-of interest detection, segmentation, and grouping. Due to these difficulties, geometric features are not widely used.
89 Image CLEF Data  ImageCLEF aims to provide an evaluation forum for the cross– language annotation and retrieval of images (launched in 2003)  Statistics of this database : – With more than 300,000 (in .JPEG format), the total size of the database > 300 GB – contains PET, CT, MRI, and Ultrasound images  Three Tasks – Modality classification – Image–based retrieval – Case–based retrieval Medical Image Database available at http://www.imageclef.org/2013/medical
90 Modality Classification Task Modality is one of the most important filters that clinicians would like to be able to limit their search by.
91 Image based and Case-based Querying  Image-based retrieval : This is the classic medical retrieval task. Similar to Query by Image Example. Given the query image, find the most similar images.  Case-based retrieval: This is a more complex task; is closer to the clinical workflow. A case description, with patient demographics, limited symptoms and test results including imaging studies, is provided (but not the final diagnosis). The goal is to retrieve cases including images that might best suit the provided case description.
92 Challenges with Image Data  Extracting informative features.  Selection of relevant features.  Sparse methods* and dimensionality reducing techniques  Integration of Image data with other data available  Early Fusion  Vector-based Integration  Intermediate Fusion  Multiple Kernel Learning  Late Fusion  Ensembling results from individual modalities *Jieping Ye’s SDM 2010 Tutorial on Sparse methods http://www.public.asu.edu/~jye02/Tutorial/Sparse-SDM10.pdf
93 Publicly Available Medical Image Repositories Image database Name Moda lities No. Of patients No. Of Images Size Of Data Notes/Applications Download Link Cancer Imaging Archive Database CT DX CR 1010 244,527 241 GB Lesion Detection and classification, Accelerated Diagnostic Image Decision, Quantitative image assessment of drug response https://public.cancerimagingarchive.net/ ncia/dataBasketDisplay.jsf Digital Mammog raphy database DX 2620 9,428 211 GB Research in Development of Computer Algorithm to aid in screening http://marathon.csee.usf.edu/Mammogr aphy/Database.html Public Lung Image Database CT 119 28,227 28 GB Identifying Lung Cancer by Screening Images https://eddie.via.cornell.edu/crpf.html Image CLEF Database PET CT MRI US unknown 306,549 316 GB Modality Classification , Visual Image Annotation , Scientific Multimedia Data Management http://www.imageclef.org/2013/medical MS Lesion Segment ation MRI 41 145 36 GB Develop and Compare 3D MS Lesion Segmentation Techniques http://www.ia.unc.edu/MSseg/download .php ADNI Database MRI PET 2851 67,871 16GB Define the progression of Alzheimer’s disease http://adni.loni.ucla.edu/data- samples/acscess-data/
94 GENETIC DATA
95 Genetic Data  The human genome is made up of DNA which consists of four different chemical building blocks (called bases and abbreviated A, T, C, and G).  It contains 3 billion pairs of bases and the particular order of As, Ts, Cs, and Gs is extremely important.  Size of a single human genome is about 3GB.  Thanks to the Human Genome Project (1990-2003) – The goal was to determine the complete sequence of the 3 billion DNA subunits (bases). – The total cost was around $3 billion.
96 Genetic Data  The whole genome sequencing data is currently being annotated and not many analytics have been applied so far since the data is relatively new.  Several publicly available genome repositories. http://aws.amazon.com/1000genomes/  It costs around $5000 to get a complete genome. It is still in the research phase. Heavily used in the cancer biology.  In this tutorial, we will focus on Genome-Wide Association Studies (GWAS). – It is more relevant to healthcare practice. Some clinical trials have already started using GWAS. – Most of the computing literature (in terms of analytics) is available for the GWAS. It is still in rudimentary stage for whole genome sequences.
97 Genome-Wide Association Studies (GWAS)  Genome-wide association studies (GWAS) are used to identify common genetic factors that influence health and disease.  These studies normally compare the DNA of two groups of participants: people with the disease (cases) and similar people without (controls). (One million Loci)  Single nucleotide polymorphisms (SNPs) are DNA sequence variations that occur when a single nucleotide (A,T,C,or G) in the genome sequence differs between individuals.  SNPs occur every 100 to 300 bases along the 3-billion-base human genome.
98 Epistasis Modeling  For simple Mendelian diseases, single SNPs can explain phenotype very well.  The complex relationship between genotype and phenotype is inadequately described by marginal effects of individual SNPs.  Increasing empirical evidence suggests that interactions among loci contribute broadly to complex traits.  The difficulty in the problem of detecting SNP pair interactions is the heavy computational burden. – To detect pairwise interactions from 500,000 SNPs genotyped in thousands of samples, a total of 1.25 X 10 statistical tests are needed.
99 Epistasis Detection Methods  Exhaustive – Enumerates all K-locus interactions among SNPs. – Efficient implementations mostly aiming at reducing computations by eliminating unnecessary calculations.  Non-Exhaustive – Stochastic: randomized search. Performance lowers when the # SNPs increase. – Heuristic: greedy methods that do not guarantee optimal solution. Shang, Junliang, et al. "Performance analysis of novel methods for detecting epistasis." BMC bioinformatics 12.1 (2011): 475.
100 Sparse Methods for SNP data analysis  Successful identification of SNPs strongly predictive of disease promises a better understanding of the biological mechanisms underlying the disease.  Sparse linear methods have been used to fit the genotype data and obtain a selected set of SNPs.  Minimizing the squared loss function (L) of N individuals and p variables (SNPs) is used for linear regression and is defined as where xi ∈ ℝp are inputs for the ith sample, y ∈ ℝN is the N vector of outputs, β0 ∈ ℝ is the intercept, β ∈ ℝp is a p-vector of model weights, and λ is user penalty.  Efficient implementations that scale to genome-wide data are available.  SparSNP package http://bioinformatics.research.nicta.com.au/software/sparsnp/ Wu, Tong Tong, et al. "Genome-wide association analysis by lasso penalized logistic regression." Bioinformatics 25.6 (2009): 714-721.     N i p j j T ii xyL 1 1 2 00 )( 2 1 ),( 
101 Public Resources for Genetic (SNP) Data  The Wellcome Trust Case Control Consortium (WTCCC) is a group of 50 research groups across the UK which was established in 2005.  Available at http://www.wtccc.org.uk/  Seven different diseases: bipolar disorder (1868 individuals), coronary heart disease (1926 individuals), Crohn's disease (1748 individuals), hypertension (1952 individuals), rheumatoid arthritis (1860 individuals), type I diabetes (1963 individuals) or type II diabetes (1924 individuals).  Around 3,000 healthy controls common for these disorders. The individuals were genotyped using Affymetrix chip and obtained approximately 500K SNPs.  The database of Genotypes and Phenotypes (dbGaP) maintained by National Center of Biotechnology Information (NCBT) at NIH.  Available at http://www.ncbi.nlm.nih.gov/gap
102 BEHAVIORAL AND PUBLIC HEALTH DATA
103 Epidemiology Data  The Surveillance Epidemiology and End Results Program (SEER) at NIH.  Publishes cancer incidence and survival data from population-based cancer registries covering approximately 28% of the population of the US.  Collected over the past 40 years (starting from January 1973 until now).  Contains a total of 7.7M cases and >350,000 cases are added each year.  Collect data on patient demographics, tumor site, tumor morphology and stage at diagnosis, first course of treatment, and follow-up for vital status. Usage:  Widely used for understanding disparities related to race, age, and gender.  Can be used to overlay information with other sources of data (such as water/air pollution, climate, socio-economic) to identify any correlations.  Can not be used for predictive analysis, but mostly used for studying trends. SEER database is available at http://seer.cancer.gov/
104 Social Media can Sense Public Health !! Chunara, Rumi, Jason R. Andrews, and John S. Brownstein. "Social and news media enable estimation of epidemiological patterns early in the 2010 Haitian cholera outbreak." American Journal of Tropical Medicine and Hygiene 86.1 (2012): 39. During infectious disease outbreaks, data collected through health institutions and official reporting structures may not be available for weeks, hindering early epidemiologic assessment. Social media can get it in near real-time. Dugas, Andrea Freyer, et al. "Google Flu Trends: correlation with emergency department influenza rates and crowding metrics." Clinical infectious diseases 54.4 (2012): 463-469. Google Flu Trends correlated with Influenza outbreak Twitter messaging correlated with cholera outbreak
105 Social Networks for Patients  PatientsLikeMe1 is a patient network is an online data sharing platform started in 2006; now has more than 200,000 patients and is tracking 1,500 diseases. 1 http://www.patientslikeme.com/ OBJECTIVE: “Given my status, what is the best outcome I can hope to achieve, and how do I get there?”  People connect with others who have the same disease or condition, track and share their own experiences, see what treatments have helped other patients like them, gain insights and identify any patterns.  Patient provides the data on their conditions, treatment history, side effects, hospitalizations, symptoms, disease-specific functional scores, weight, mood, quality of life and more on an ongoing basis.  Gaining access to the patients for future clinical trials.
106 Home Monitoring and Sensing Technologies  Advancements in sensing technology are critical for developing effective and efficient home-monitoring systems • Sensing devices can provide several types of data in real-time. • Activity Recognition using Cell Phone Accelerometers Kwapisz, Jennifer R., Gary M. Weiss, and Samuel A. Moore. "Activity recognition using cell phone accelerometers." ACM SIGKDD Explorations Newsletter 12.2 (2011): 74-82. Rashidi, Parisa, et al. "Discovering activities to recognize and track in a smart environment." Knowledge and Data Engineering, IEEE Transactions on 23.4 (2011): 527-539.
107 Public Health and Behavior Data Repositories Dataset Link Description Behavioral Risk Factor Surveillance System (BRFSS) http://www.cdc.gov/brfss/technical_ infodata/index.htm Healthcare survey data: smoking, alcohol, lifestyle (diet, exercise), major diseases (diabetes, cancer), mental illness Ohio Hospital Inpatient/Outpatient Data http://publicapps.odh.ohio.gov/pwh /PWHMain.aspx?q=021813114232 Hospital: number of discharges, transfers, length of stay, admissions, transfers, number of patients with specific procedure codes US Mortality Data http://www.cdc.gov/nchs/data_acce ss/cmf.htm Mortality information on county- level Human Mortality Database http://www.mortality.org/ Birth, death, population size by country Utah Public Health Database http://ibis.health.utah.gov/query Summary statistics for mortality, charges, discharges, length of stay on a county-level basis Dartmouth Atlas of Health Care http://www.dartmouthatlas.org/tools /downloads.aspx Post discharge events, chronically ill care, surgical discharge rate Thanks to Prof. Joydeep Ghosh from UT Austin for providing this information.
108 CONCLUDING REMARKS
109 Final Thoughts Big data could save the health care industry up to $450 billion, but other things are important too.  Right living: Patients should take more active steps to improve their health.  Right care: Developing a coordinated approach to care in which all caregivers have access to the same information.  Right provider: Any professionals who treat patients must have strong performance records and be capable of achieving the best outcomes.  Right value: Improving value while simultaneously improving care quality.  Right innovation: Identifying new approaches to health-care delivery. “Stakeholders will only benefit from big data if they take a more holistic, patient-centered approach to value, one that focuses equally on health-care spending and treatment outcomes,” McKinsey report available at: http://www.mckinsey.com/insights/health_systems/the_big-data_revolution_in_us_health_care
110 Conclusion  Big data analytics is a promising right direction which is in its infancy for the healthcare domain.  Healthcare is a data-rich domain. As more and more data is being collected, there will be increasing demand for big data analytics.  Unraveling the “Big Data” related complexities can provide many insights about making the right decisions at the right time for the patients.  Efficiently utilizing the colossal healthcare data repositories can yield some immediate returns in terms of patient outcomes and lowering care costs.  Data with more complexities keep evolving in healthcare thus leading to more opportunities for big data analytics.
111 Acknowledgements • Funding Sources – National Science foundation – National Institutes of Health – Susan G. Komen for the Cure – Delphinus Medical Technologies – IBM Research
112 Thank You Questions and Comments Feel free to email questions or suggestions to jimeng@cs.cmu.edu reddy@cs.wayne.edu

More Related Content

PPTX
Deep-learning-or-health-informatics-recent-trends-and-future-directions By Ra...
byraihansikdar
 
PDF
IBM Watson in Healthcare
byAnders Quitzau
 
PPTX
Role of Big Data in Medical Diagnostics
byNishant Agarwal
 
PDF
IBM Watson Health: How cognitive technologies have begun transforming clinica...
byMaged N. Kamel Boulos
 
PDF
IBM Watson for Healthcare
byIBM_CH
 
PDF
Rise of the Machines
byHeather Fraser
 
PDF
IRJET- Integration of Big Data Analytics in Healthcare Systems
byIRJET Journal
 
PDF
Mie2014 workshop: Gap Analysis of Personalized Health Services through Patien...
byPei-Yun Sabrina Hsueh
 
Deep-learning-or-health-informatics-recent-trends-and-future-directions By Ra...
byraihansikdar
 
IBM Watson in Healthcare
byAnders Quitzau
 
Role of Big Data in Medical Diagnostics
byNishant Agarwal
 
IBM Watson Health: How cognitive technologies have begun transforming clinica...
byMaged N. Kamel Boulos
 
IBM Watson for Healthcare
byIBM_CH
 
Rise of the Machines
byHeather Fraser
 
IRJET- Integration of Big Data Analytics in Healthcare Systems
byIRJET Journal
 
Mie2014 workshop: Gap Analysis of Personalized Health Services through Patien...
byPei-Yun Sabrina Hsueh
 

What's hot

PDF
HEC 2016 Panel: Putting User-Generated Data in Action: Improving Interpretabi...
byPei-Yun Sabrina Hsueh
 
PDF
Big data for health
byredpel dot com
 
PDF
Healthcare Innovation and Transformation - Dr. Ken Yale
byKen Yale
 
PDF
Deep Learning in Medicine
byijtsrd
 
PPTX
Medical Question Answering: Dealing with the complexity and specificity of co...
byAsma Ben Abacha
 
PPTX
The Hive Think Tank: Unpacking AI for Healthcare
byThe Hive
 
PDF
The Road toward a Smart Hospital (Presented at Roi Et Hospital) (2 Feb 2016)
byNawanan Theera-Ampornpunt
 
PPTX
Insights from the Organization of International Challenges on Artificial Inte...
byAsma Ben Abacha
 
PDF
The Future of Medical Education From Dreams to Reality (VR, AR, AI)
bySeriousGamesAssoc
 
PPTX
Multimodal Question Answering in the Medical Domain (CMU/LTI 2020) | Dr. Asma...
byAsma Ben Abacha
 
PPTX
The future of healthcare and big data
byCharles Barnett
 
PPTX
Developing a World Leading Technology Enabled Health Programme of Research
byMaged N. Kamel Boulos
 
DOCX
Precision medicine and AI: problems ahead
byNeil Raden
 
DOCX
Tim Capstone Paper(2)
byTimothy Greiner
 
PPTX
Big Data and Smart Healthcare
bySujan Perera
 
PDF
IBM Terkko Pop-up Presentation by Pekka Leppänen
byTerkkoHub
 
PDF
ICT in Healthcare: Thailand 4.0
byNawanan Theera-Ampornpunt
 
PPTX
Watson – from Jeopardy to healthcare
byHealth Informatics New Zealand
 
PPTX
Augmented Personalized Health: an explicit knowledge enhanced neurosymbolic d...
byAmit Sheth
 
PDF
New Normal, New Future - Free Download E book
bykevin brown
 
HEC 2016 Panel: Putting User-Generated Data in Action: Improving Interpretabi...
byPei-Yun Sabrina Hsueh
 
Big data for health
byredpel dot com
 
Healthcare Innovation and Transformation - Dr. Ken Yale
byKen Yale
 
Deep Learning in Medicine
byijtsrd
 
Medical Question Answering: Dealing with the complexity and specificity of co...
byAsma Ben Abacha
 
The Hive Think Tank: Unpacking AI for Healthcare
byThe Hive
 
The Road toward a Smart Hospital (Presented at Roi Et Hospital) (2 Feb 2016)
byNawanan Theera-Ampornpunt
 
Insights from the Organization of International Challenges on Artificial Inte...
byAsma Ben Abacha
 
The Future of Medical Education From Dreams to Reality (VR, AR, AI)
bySeriousGamesAssoc
 
Multimodal Question Answering in the Medical Domain (CMU/LTI 2020) | Dr. Asma...
byAsma Ben Abacha
 
The future of healthcare and big data
byCharles Barnett
 
Developing a World Leading Technology Enabled Health Programme of Research
byMaged N. Kamel Boulos
 
Precision medicine and AI: problems ahead
byNeil Raden
 
Tim Capstone Paper(2)
byTimothy Greiner
 
Big Data and Smart Healthcare
bySujan Perera
 
IBM Terkko Pop-up Presentation by Pekka Leppänen
byTerkkoHub
 
ICT in Healthcare: Thailand 4.0
byNawanan Theera-Ampornpunt
 
Watson – from Jeopardy to healthcare
byHealth Informatics New Zealand
 
Augmented Personalized Health: an explicit knowledge enhanced neurosymbolic d...
byAmit Sheth
 
New Normal, New Future - Free Download E book
bykevin brown
 

Viewers also liked

PDF
Chapter 2 research methodlogy
byAravindharamanan S
 
PPT
Introduction to recommendation system
byAravindharamanan S
 
PPT
Cs548 s15 showcase_web_mining
byAravindharamanan S
 
PPT
Content based recommendation systems
byAravindharamanan S
 
PPT
Combining content based and collaborative filtering
byAravindharamanan S
 
PDF
Big data-analytics-2013-peer-research-report
byAravindharamanan S
 
PDF
Full xml
byAravindharamanan S
 
PDF
uae views on big data
byAravindharamanan S
 
PDF
Tomcat + other things
byAravindharamanan S
 
DOCX
Database development connection steps
byAravindharamanan S
 
PDF
Android chapter18 c-internet-web-services
byAravindharamanan S
 
PPT
Personalizing the web building effective recommender systems
byAravindharamanan S
 
PDF
Rest hello world_tutorial
byAravindharamanan S
 
PDF
Android ui layouts ,cntls,webservices examples codes
byAravindharamanan S
 
PDF
Sql developer usermanual_en
byAravindharamanan S
 
Chapter 2 research methodlogy
byAravindharamanan S
 
Introduction to recommendation system
byAravindharamanan S
 
Cs548 s15 showcase_web_mining
byAravindharamanan S
 
Content based recommendation systems
byAravindharamanan S
 
Combining content based and collaborative filtering
byAravindharamanan S
 
Big data-analytics-2013-peer-research-report
byAravindharamanan S
 
Full xml
byAravindharamanan S
 
uae views on big data
byAravindharamanan S
 
Tomcat + other things
byAravindharamanan S
 
Database development connection steps
byAravindharamanan S
 
Android chapter18 c-internet-web-services
byAravindharamanan S
 
Personalizing the web building effective recommender systems
byAravindharamanan S
 
Rest hello world_tutorial
byAravindharamanan S
 
Android ui layouts ,cntls,webservices examples codes
byAravindharamanan S
 
Sql developer usermanual_en
byAravindharamanan S
 

Similar to Sun==big data analytics for health care

PPTX
Use of data analytics in health care
byAkanshabhushan
 
PDF
eBook - Data Analytics in Healthcare
byNextGen Healthcare
 
PDF
A Systematic Review Of Type-2 Diabetes By Hadoop Map-Reduce
byFinni Rice
 
PDF
Big implications of Big Data in healthcare
byGuires
 
PPTX
Applications of Data Science in Healthcare
byBrijitSaraJoseph
 
PPTX
Big data analytics in healthcare
byJoseph Thottungal
 
PDF
Rock Report: Big Data by @Rock_Health
byRock Health
 
PPTX
Big data analystics
byAll India Institute of Medical Sciences
 
PDF
Big Data Solutions for Healthcare
byOdinot Stanislas
 
DOCX
Ajith M Jose_Report1.docx
bymca2206
 
PDF
Big Data and its Impact on Industry (Example of the Pharmaceutical Industry)
byHellmuth Broda
 
PPTX
McGrath Health Data Analyst SXSW
byRobert McGrath
 
PDF
Review paper on Big Data in healthcare informatics
byIRJET Journal
 
PPTX
Designingforinnovationppayneucsf06 2105final-150610135139-lva1-app6892
byRARE-data = Solutions d/b/a RARE Advisors, LLC
 
PPTX
HEALTHCARE_IT
byAmshumala S
 
PPT
JR's Lifetime Advanced Analytics
byd-Wise Technologies
 
PPT
JR's Lifetime Advanced Analytics
byChase Hamilton
 
PDF
IRJET- A Survey on Big Data Frameworks and Approaches in Health Care Sector
byIRJET Journal
 
PPTX
Optimum Healthcare IT A physician’s perspective on Big Data, Predictive Analy...
byHealthXn
 
PPTX
E.Gombocz: Changing the Model in Pharma and Healthcare (DILS Keynote 2013-07...
byErich Gombocz
 
Use of data analytics in health care
byAkanshabhushan
 
eBook - Data Analytics in Healthcare
byNextGen Healthcare
 
A Systematic Review Of Type-2 Diabetes By Hadoop Map-Reduce
byFinni Rice
 
Big implications of Big Data in healthcare
byGuires
 
Applications of Data Science in Healthcare
byBrijitSaraJoseph
 
Big data analytics in healthcare
byJoseph Thottungal
 
Rock Report: Big Data by @Rock_Health
byRock Health
 
Big data analystics
byAll India Institute of Medical Sciences
 
Big Data Solutions for Healthcare
byOdinot Stanislas
 
Ajith M Jose_Report1.docx
bymca2206
 
Big Data and its Impact on Industry (Example of the Pharmaceutical Industry)
byHellmuth Broda
 
McGrath Health Data Analyst SXSW
byRobert McGrath
 
Review paper on Big Data in healthcare informatics
byIRJET Journal
 
Designingforinnovationppayneucsf06 2105final-150610135139-lva1-app6892
byRARE-data = Solutions d/b/a RARE Advisors, LLC
 
HEALTHCARE_IT
byAmshumala S
 
JR's Lifetime Advanced Analytics
byd-Wise Technologies
 
JR's Lifetime Advanced Analytics
byChase Hamilton
 
IRJET- A Survey on Big Data Frameworks and Approaches in Health Care Sector
byIRJET Journal
 
Optimum Healthcare IT A physician’s perspective on Big Data, Predictive Analy...
byHealthXn
 
E.Gombocz: Changing the Model in Pharma and Healthcare (DILS Keynote 2013-07...
byErich Gombocz
 

Recently uploaded

PPTX
NURSING PROCESS.....................pptx
byAneetaSharma15
 
PDF
Unit - I Communication Skills. B.pharmacy and D. Pharmacy
bySayyadTarannum
 
PDF
JRF Plant Science Preparation Strategy by Experts | Complete Syllabus Discuss...
bySatyam Sharma
 
PDF
Projects_for_Ministries,_Departments_&_Agencies2123341679.pdf
bynservice241
 
PDF
JRF Plant Science Syllabus | Complete Unit-Wise Discussion & Preparation Stra...
bySatyam Sharma
 
PPTX
Nasopulmonary drug delivery system Pradip Sontakke
byDrxpradipPatil
 
PPTX
Warehouse in Traceability Report - Odoo 19
byCeline George
 
PPTX
What are Forecast Reports in Odoo 18 CRM
byCeline George
 
PPTX
Overview of Uploading a bill in Odoo 19
byCeline George
 
PPTX
DepEd_ARAL_Program_Orientation FINAL COPY.pptx
byAljohnEspejo1
 
PPTX
GASTROINTESTINAL DISORDERS(1-Upper Gastrointestinal Bleeding 2-Lower Gastroi...
byDr PANKAJ PARMAR
 
PPTX
Overview of Cash on delivery in Odoo 19
byCeline George
 
PPTX
Unit 3-proteins.pptx....................
byAkanksha Kotangale
 
PDF
2025-11-02 - AQSE - Qiskit Fall Fest - IIT-Guwahati.pdf
byAritra Sarkar
 
PDF
Hunwick's Roman Roads and Ridge Roads, a Discussion Document
byJohn Pallister
 
PPTX
An inclusive AI - Supporting students with special education needs by Andreas...
byEduSkills OECD
 
PDF
BP704T: NOVEL DRUG DELIVERY SYSTEMS UNIT 4 PART 1
byRushiMandali
 
PDF
BP605 T PHARMACEUTICAL BIOTECHNOLOGY UNIT 4 PART 1
byRushiMandali
 
PDF
2025 Caribbean Examinations Council CCSLC Merit List
byCaribbean Examinations Council
 
PDF
2025 Caribbean Examinations Council CAPE Merit List
byCaribbean Examinations Council
 
NURSING PROCESS.....................pptx
byAneetaSharma15
 
Unit - I Communication Skills. B.pharmacy and D. Pharmacy
bySayyadTarannum
 
JRF Plant Science Preparation Strategy by Experts | Complete Syllabus Discuss...
bySatyam Sharma
 
Projects_for_Ministries,_Departments_&_Agencies2123341679.pdf
bynservice241
 
JRF Plant Science Syllabus | Complete Unit-Wise Discussion & Preparation Stra...
bySatyam Sharma
 
Nasopulmonary drug delivery system Pradip Sontakke
byDrxpradipPatil
 
Warehouse in Traceability Report - Odoo 19
byCeline George
 
What are Forecast Reports in Odoo 18 CRM
byCeline George
 
Overview of Uploading a bill in Odoo 19
byCeline George
 
DepEd_ARAL_Program_Orientation FINAL COPY.pptx
byAljohnEspejo1
 
GASTROINTESTINAL DISORDERS(1-Upper Gastrointestinal Bleeding 2-Lower Gastroi...
byDr PANKAJ PARMAR
 
Overview of Cash on delivery in Odoo 19
byCeline George
 
Unit 3-proteins.pptx....................
byAkanksha Kotangale
 
2025-11-02 - AQSE - Qiskit Fall Fest - IIT-Guwahati.pdf
byAritra Sarkar
 
Hunwick's Roman Roads and Ridge Roads, a Discussion Document
byJohn Pallister
 
An inclusive AI - Supporting students with special education needs by Andreas...
byEduSkills OECD
 
BP704T: NOVEL DRUG DELIVERY SYSTEMS UNIT 4 PART 1
byRushiMandali
 
BP605 T PHARMACEUTICAL BIOTECHNOLOGY UNIT 4 PART 1
byRushiMandali
 
2025 Caribbean Examinations Council CCSLC Merit List
byCaribbean Examinations Council
 
2025 Caribbean Examinations Council CAPE Merit List
byCaribbean Examinations Council
 

Sun==big data analytics for health care

  • 1.
    1 Big Data Analyticsfor Healthcare Chandan K. Reddy Department of Computer Science Wayne State University Tutorial presentation at the SIAM International Conference on Data Mining, Austin, TX, 2013. The updated tutorial slides are available at http://dmkd.cs.wayne.edu/TUTORIAL/Healthcare/ Jimeng Sun Healthcare Analytics Department IBM TJ Watson Research Center
  • 2.
    2 Motivation Hersh, W., Jacko,J. A., Greenes, R., Tan, J., Janies, D., Embi, P. J., & Payne, P. R. (2011). Health-care hit or miss? Nature, 470(7334), 327. In 2012, worldwide digital healthcare data was estimated to be equal to 500 petabytes and is expected to reach 25,000 petabytes in 2020. Can we learn from the past to become better in the future ?? Healthcare Data is becoming more complex !!
  • 3.
    3 Organization of thisTutorial  Introduction  Motivating Examples  Sources and Techniques for Big Data in Healthcare – Structured EHR Data – Unstructured Clinical Notes – Medical Imaging Data – Genetic Data – Other Data (Epidemiology & Behavioral)  Final Thoughts and Conclusion
  • 4.
  • 5.
    5 Definition of BigData  A collection of large and complex data sets which are difficult to process using common database management tools or traditional data processing applications. The challenges include capturing, storing, searching, sharing & analyzing. The four dimensions (V’s) of Big Data Big data is not just about size. • Finds insights from complex, noisy, heterogeneous, longitudinal, and voluminous data. • It aims to answer questions that were previously unanswered. BIG DATA Velocity Veracity Variety Volume  “Big data refers to the tools, processes and procedures allowing an organization to create, manipulate, and manage very large data sets and storage facilities” – according to zdnet.com
  • 6.
    6 Reasons for GrowingComplexity/Abundance of Healthcare Data  Standard medical practice is moving from relatively ad-hoc and subjective decision making to evidence-based healthcare.  More incentives to professionals/hospitals to use EHR technology. Additional Data Sources  Development of new technologies such as capturing devices, sensors, and mobile applications.  Collection of genomic information became cheaper.  Patient social communications in digital forms are increasing.  More medical knowledge/discoveries are being accumulated.
  • 7.
    7 Big Data Challengesin Healthcare  Inferring knowledge from complex heterogeneous patient sources. Leveraging the patient/data correlations in longitudinal records.  Understanding unstructured clinical notes in the right context.  Efficiently handling large volumes of medical imaging data and extracting potentially useful information and biomarkers.  Analyzing genomic data is a computationally intensive task and combining with standard clinical data adds additional layers of complexity.  Capturing the patient’s behavioral data through several sensors; their various social interactions and communications.
  • 8.
    8  Take advantageof the massive amounts of data and provide right intervention to the right patient at the right time.  Personalized care to the patient.  Potentially benefit all the components of a healthcare system i.e., provider, payer, patient, and management. Electronic Health Records Evidence + Insights Improved outcomes through smarter decisions Lower costs Big Data Analytics Overall Goals of Big Data Analytics in Healthcare BehavioralGenomic Public Health
  • 9.
    9 Purpose of thisTutorial Two-fold objectives:  Introduce the data mining researchers to the sources available and the possible challenges and techniques associated with using big data in healthcare domain.  Introduce Healthcare analysts and practitioners to the advancements in the computing field to effectively handle and make inferences from voluminous and heterogeneous healthcare data. The ultimate goal is to bridge data mining and medical informatics communities to foster interdisciplinary works between the two communities. PS: Due to the broad nature of the topic, the primary emphasis will be on introducing healthcare data repositories, challenges, and concepts to data scientists. Not much focus will be on describing the details of any particular techniques and/or solutions.
  • 10.
    10 Disclaimers  Being arecent and growing topic, there might be several other resources that might not be covered here.  Presentation here is more biased towards the data scientists’ perspective and may be less towards the healthcare management or healthcare provider’s perspective.  Some of the website links provided might become obsolete in the future. This tutorial is prepared in early 2013.  Since this topic contains a wide varieties of problems, there might be some aspects of healthcare that might not be covered in the tutorial.
  • 11.
  • 12.
    12 EXAMPLE 1: HeritageHealth Prize  Over $30 billion was spent on unnecessary hospital admissions. Goals:  Identify patients at high-risk and ensure they get the treatment they need.  Develop algorithms to predict the number of days a patient will spend in a hospital in the next year. Outcomes:  Health care providers can develop new strategies to care for patients before its too late reduces the number of unnecessary hospitalizations.  Improving the health of patients while decreasing the costs of care.  Winning solutions use a combination of several predictive models. http://www.heritagehealthprize.com
  • 13.
    13 EXAMPLE 2: Penaltiesfor Poor Care - 30-Day Readmissions  Hospitalizations account for more than 30% of the 2 trillion annual cost of healthcare in the United States. Around 20% of all hospital admissions occur within 30 days of a previous discharge. – not only expensive but are also potentially harmful, and most importantly, they are often preventable.  Medicare penalizes hospitals that have high rates of readmissions among patients with heart failure, heart attack, and pneumonia.  Identifying patients at risk of readmission can guide efficient resource utilization and can potentially save millions of healthcare dollars each year.  Effectively making predictions from such complex hospitalization data will require the development of novel advanced analytical models.
  • 14.
    14 EXAMPE 3: WhiteHouse unveils BRAIN Initiative  The US President unveiled a new bold $100 million research initiative designed to revolutionize our understanding of the human brain. BRAIN (Brain Research through Advancing Innovative Neurotechnologies) Initiative.  Find new ways to treat, cure, and even prevent brain disorders, such as Alzheimer’s disease, epilepsy, and traumatic brain injury.  “Every dollar we invested to map the human genome returned $140 to our economy... Today, our scientists are mapping the human brain to unlock the answers to Alzheimer’s.” -- President Barack Obama, 2013 State of the Union.  “advances in "Big Data" that are necessary to analyze the huge amounts of information that will be generated; and increased understanding of how thoughts, emotions, actions and memories are represented in the brain.” : NSF  Joint effort by NSF, NIH, DARPA, and other private partners. http://www.whitehouse.gov/infographics/brain-initiative
  • 15.
    15 EXAMPLE 4: GEHead Health Challenge Challenge 1: Methods for Diagnosis and Prognosis of Mild Traumatic Brain Injuries. Challenge 2: The Mechanics of Injury: Innovative Approaches For Preventing And Identifying Brain Injuries. In Challenge 1, GE and the NFL will award up to $10M for two types of solutions: Algorithms and Analytical Tools, and Biomarkers and other technologies. A total of $60M in funding over a period of 4 years.
  • 16.
    16 Healthcare Continuum Sarkar, IndraNeil. "Biomedical informatics and translational medicine." Journal of Translational Medicine 8.1 (2010): 22.
  • 17.
    17 Data Collection andAnalysis Jensen, Peter B., Lars J. Jensen, and Søren Brunak. "Mining electronic health records: towards better research applications and clinical care." Nature Reviews Genetics (2012). Effectively integrating and efficiently analyzing various forms of healthcare data over a period of time can answer many of the impending healthcare problems.
  • 18.
    18 Organization of thisTutorial  Introduction  Motivating Examples  Sources and Techniques for Big Data in Healthcare – Structured EHR Data – Unstructured Clinical Notes – Medical Imaging Data – Genetic Data – Other Data (Epidemiology & Behavioral)  Final Thoughts and Conclusion
  • 19.
    19 SOURCES AND TECHNIQUES FORBIG DATA IN HEALTHCARE
  • 20.
    20 Outline Electronic Health Records(EHR) data Healthcare Analytic Platform Resources
  • 21.
  • 22.
    22 Data Health data Genomic data • DNAsequences Clinical data • Structured EHR • Unstructured EHR • Medical Images Behavior data • Social network data • Mobility sensor data
  • 23.
    23 Billing data -ICD codes  ICD stands for International Classification of Diseases  ICD is a hierarchical terminology of diseases, signs, symptoms, and procedure codes maintained by the World Health Organization (WHO)  In US, most people use ICD-9, and the rest of world use ICD-10  Pros: Universally available  Cons: medium recall and medium precision for characterizing patients • (250) Diabetes mellitus • (250.0) Diabetes mellitus without mention of complication • (250.1) Diabetes with ketoacidosis • (250.2) Diabetes with hyperosmolarity • (250.3) Diabetes with other coma • (250.4) Diabetes with renal manifestations • (250.5) Diabetes with ophthalmic manifestations • (250.6) Diabetes with neurological manifestations • (250.7) Diabetes with peripheral circulatory disorders • (250.8) Diabetes with other specified manifestations • (250.9) Diabetes with unspecified complication
  • 24.
    24 Billing data –CPT codes  CPT stands for Current Procedural Terminology created by the American Medical Association  CPT is used for billing purposes for clinical services  Pros: High precision  Cons: Low recall Codes for Evaluation and Management: 99201-99499 (99201 - 99215) office/other outpatient services (99217 - 99220) hospital observation services (99221 - 99239) hospital inpatient services (99241 - 99255) consultations (99281 - 99288) emergency dept services (99291 - 99292) critical care services …
  • 25.
    25 Lab results  Thestandard code for lab is Logical Observation Identifiers Names and Codes (LOINC®)  Challenges for lab – Many lab systems still use local dictionaries to encode labs – Diverse numeric scales on different labs • Often need to map to normal, low or high ranges in order to be useful for analytics – Missing data • not all patients have all labs • The order of a lab test can be predictive, for example, BNP indicates high likelihood of heart failure Time Lab Value 1996-03-15 12:50:00.0 CO2 29.0 1996-03-15 12:50:00.0 BUN 16.0 1996-03-15 12:50:00.0 HDL-C 37.0 1996-03-15 12:50:00.0 K 4.5 1996-03-15 12:50:00.0 Cl 102.0 1996-03-15 12:50:00.0 Gluc 86.0
  • 26.
    26 Medication  Standard codeis National Drug Code (NDC) by Food and Drug Administration (FDA), which gives a unique identifier for each drug – Not used universally by EHR systems – Too specific, drugs with the same ingredients but different brands have different NDC  RxNorm: a normalized naming system for generic and branded drugs by National Library of Medicine  Medication data can vary in EHR systems – can be in both structured or unstructured forms  Availability and completeness of medication data vary – Inpatient medication data are complete, but outpatient medication data are not – Medication usually only store prescriptions but we are not sure whether patients actually filled those prescriptions
  • 27.
    27 Clinical notes  Clinicalnotes contain rich and diverse source of information  Challenges for handling clinical notes – Ungrammatical, short phrases – Abbreviations – Misspellings – Semi-structured information • Copy-paste from other structure source – Lab results, vital signs • Structured template: – SOAP notes: Subjective, Objective, Assessment, Plan
  • 28.
    28 Summary of commonEHR data ICD CPT Lab Medication Clinical notes Availability High High High Medium Medium Recall Medium Poor Medium Inpatient: High Outpatient: Variable Medium Precision Medium High High Inpatient: High Outpatient: Variable Medium high Format Structured Structured Mostly structured Structured and unstructured Unstructured Pros Easy to work with, a good approximation of disease status Easy to work with, high precision High data validity High data validity More details about doctors’ thoughts Cons Disease code often used for screening, therefore disease might not be there Missing data Data normalization and ranges Prescribed not necessary taken Difficult to process Joshua C. Denny Chapter 13: Mining Electronic Health Records in the Genomics Era. PLoS Comput Biol. 2012 December; 8(12):
  • 29.
  • 30.
  • 31.
  • 32.
  • 33.
  • 34.
    34 Text Mining inHealthcare  Text mining – Information Extraction • Name Entity Recognition – Information Retrieval  Clinical text vs. Biomedical text – Biomedical text: medical literatures (well-written medical text) – Clinical text is written by clinicians in the clinical settings • Meystre et al. Extracting Information from Textual Documents in the Electronic Health Record: A Review of Recent Research. IMIA 2008 • Zweigenbaum et al. Frontiers of biomedical text mining: current progress, BRIEFINGS IN BIOINFORMATICS. VOL 8. NO 5. 358-375 • Cohen and Hersh, A survey of current work in biomedical text mining. BRIEFINGS IN BIOINFORMATICS. VOL 6. NO 1. 57–71.
  • 35.
    35 Auto-Coding: Extracting Codesfrom Clinical Text  Problem – Automatically assign diagnosis codes to clinical text  Significance – The cost is approximately $25 billion per year in the US  Available Data – Medical NLP Challenges from 2007 • Subsections from radiology reports: clinical history and impression  Potential Evaluation Metric: – F-measure = 2P*R/(P+R), where P is precision, and R is recall.  Example References • Aronson et al. From indexing the biomedical literature to coding clinical text: experience with MTI and machine learning approaches. BioNLP 2007 • Crammer et al. Automatic Code Assignment to Medical Text. BioNLP 2007 • Friedman C, Shagina L, Lussier Y, Hripcsak G. Automated encoding of clinical documents based on natural language processing. JAMIA. 2004:392-402 • Pakhomov SV, Buntrock JD, Chute CG. Automating the assignment of diagnosis codes to patient encounters using example-based and machine learning techniques. JAMIA 2006:516-25.
  • 36.
    36 Context Analysis -Negation  Negation: e.g., ...denies chest pain… – NegExpander [1] achieves 93% precision on mammographic reports – NegEx [2] uses regular expression and achieves 94.5% specificity and 77.8% sensitivity – NegFinder [3] uses UMLS and regular expression, and achieves 97.7 specificity and 95.3% sensitivity when analyzing surgical notes and discharge summaries – A hybrid approach [4] uses regular expression and grammatical parsing and achieves 92.6% sensitivity and 99.8% specificity 1. Aronow DB, Fangfang F, Croft WB. Ad hoc classification of radiology reports. JAMIA 1999:393-411 2. Chapman et al. A simple algorithm for identifying negated findings and diseases in discharge summaries. JBI 2001:301-10. 3. Mutalik PG, et al. Use of general-purpose negation detection to augment concept indexing of medical documents: a quantitative study using the UMLS. JAMIA 2001:598-609. 4. Huang Y, Lowe HJ. A novel hybrid approach to automated negation detection in clinical radiology reports. JAMIA 2007
  • 37.
    37 Context Analysis -Temporality  Temporality: e.g., …fracture of the tibia 2 years ago – TimeText [1] can detect temporal relations with 93.2% recall and 96.9% precision on 14 discharge summaries – Context [2] is an extension of NegEx, which identifies • negations (negated, affirmed), • temporality (historical, recent, hypothetical) • experiencer (patient, other) 1. Zhou et al. The Evaluation of a Temporal Reasoning System in Processing Clinical Discharge Summaries. JAMIA 2007. 2. Chapman W, Chu D, Dowling JN. ConText: An Algorithm for Identifying Contextual Features from Clinical Text. BioNLP 2007 Trigger TerminationClinical concepts Scope of the context v v v v
  • 38.
    38 CASE 1: CASEBASED RETRIEVAL Sondhi P, Sun J, Zhai C, Sorrentino R, Kohn MS. Leveraging medical thesauri and physician feedback for improving medical literature retrieval for case queries. JAMIA. 2012
  • 39.
    39 Input: Case Query Patientwith smoking habit and weight loss. The frontal and lateral chest X rays show a mass in the posterior segment of the right upper lobe as well as a right hilar enlargement and obliteration of the right paratracheal stripe. On the chest CT the contours of the mass are lobulated with heterogeneous enhancement.Enlarged mediastinal and hilar lymph nodes are present.
  • 40.
    40 Goal: Find RelevantResearch Articles to a Query Additional Related Information Disease MeSH
  • 41.
    41 Challenge 1: QueryWeighing Patient with smoking habit and weight loss. The frontal and lateral chest X rays show a mass in the posterior segment of the right upper lobe as well as a right hilar enlargement and obliteration of the right paratracheal stripe. On the chest CT the contours of the mass are lobulated with heterogeneous enhancement. Enlarged mediastinal and hilar lymph nodes are present.  Queries are long  Not all words useful  IDF does not reflect importance  Semantics decide weight
  • 42.
    42 Method: Semantic QueryWeighing Identify important UMLS Semantic Types based on their definition Assign higher weights to query words under these types Disease or syndrome, Body part organ or organ component, Sign or symptom, Finding, Acquired abnormality, Congenital abnormality, Mental or behavioral dysfunction, Neoplasm, Pharmacologic substance, Individual Behavior Included UMLS semantic types
  • 43.
    43  Matching variants “x ray”, “x-rays”, “x rays”  Matching synonyms  “CT” or “x rays”  Knowledge gap Challenge 2: Vocabulary Gap Patient with smoking habit and weight loss. The frontal and lateral chest X rays show a mass in the posterior segment of the right upper lobe as well as a right hilar enlargement and obliteration of the right paratracheal stripe. On the chest CT the contours of the mass are lobulated with heterogeneous enhancement.Enlarged mediastinal and hilar lymph nodes are present.
  • 44.
    44 Method: Additional QueryKeywords  Asked physicians to provide additional keywords  Adding them with low weight helps  Any potential diagnosis keywords help greatly  Gives us insights into better query formulation Female patient, 25 years old, with fatigue and a swallowing disorder (dysphagia worsening during a meal). The frontal chest X-ray shows opacity with clear contours in contact with the right heart border. Right hilar structures are visible through the mass. The lateral X-ray confirms the presence of a mass in the anterior mediastinum. On CT images, the mass has a relatively homogeneous tissue density. Additional keywords: Thymoma, Lymphoma, Dysphagia, Esophageal obstruction, Myasthenia gravis, Fatiguability, Ptosis
  • 45.
    45 Challenge 3: Pseudo-Feedback Generalvocabulary gap solution: – Apply Pseudo-Relevance Feedback What if very few of the top N are relevant? No idea which keywords to pick up
  • 46.
    46 Method: Medical SubjectHeading (MeSH) Feedback Any case related query usually relates only to handful of conditions How to guess the condition of the query? –Select MeSH terms from top N=10 ranked documents –Select MeSH terms covering most query keywords –Use them for feedback
  • 47.
    47 Method: MeSH Feedback LungNeoplasmsDoc 1 BronchitisDoc 2 Cystic FibrosisDoc 3 Lung NeoplasmsDoc 4 HepatitisDoc 5
  • 48.
    48 Method 2: MeSHFeedback Lung NeoplasmsDoc 1 BronchitisDoc 2 Cystic FibrosisDoc 3 Lung NeoplasmsDoc 4 HepatitisDoc 5 Filtration List Lung Neoplasms Bronchitis
  • 49.
    49 Method 2: MeSHFeedback Lung NeoplasmsDoc 1 BronchitisDoc 2 Cystic FibrosisDoc 3 Lung NeoplasmsDoc 4 HepatitisDoc 5 Filtration List Lung Neoplasms Bronchitis Reduce Weight Leave Unchanged Reduce Weight
  • 50.
    50 Method: MeSH Feedback LungNeoplasmsDoc 1 BronchitisDoc 2 Cystic FibrosisDoc 3 Lung NeoplasmsDoc 4 HepatitisDoc 5 Filtration List Lung Neoplasms Bronchitis Reduce Weight Leave Unchanged Reduce Weight Cystic FibrosisDoc 3 Lung NeoplasmsDoc 4 HepatitisDoc 5 Lung NeoplasmsDoc 1 BronchitisDoc 2
  • 51.
    51 Retrieval Results Best performingrun Data + Knowledge helps!
  • 52.
    52 CASE 2: HEARTFAILURE SIGNS AND SYMPTOMS Roy J. Byrd, Steven R. Steinhubl, Jimeng Sun, Shahram Ebadollahi, Walter F. Stewart. Automatic identification of heart failure diagnostic criteria, using text analysis of clinical notes from electronic health records. International Journal of Medical Informatics 2013
  • 53.
    53 Framingham HF Signsand Symptoms  Framingham criteria for HF* are signs and symptoms that are documented even at primary care visits * McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: the Framingham study. N Engl J Med. 1971;285(26):1441-6.
  • 54.
    54 Natural Language Processing(NLP) Pipeline  Criteria extraction comes from sentence level.  Encounter label comes from the entire note. Roy J. Byrd, Steven R. Steinhubl, Jimeng Sun, Shahram Ebadollahi, Walter F. Stewart. Automatic identification of heart failure diagnostic criteria, using text analysis of clinical notes from electronic health records. International Journal of Medical Informatics 2013
  • 55.
    55 Performance on EncounterLevel on Test Set  Machine learning method: decision tree  Rule-based method is to construct grammars by computational linguists  Manually constructed rules are more accurate but more effort to construct than automatic rules from learning a decision tree 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Recall Precision F-Score Recall Precision F-Score Machine-learning method Rule-based method Overall Affir m e d Denied
  • 56.
    56 Potential Impact onEvidence-based Therapies 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% 70.00% Preceding Framingham diagnosis After Framingham diagnosis After clinical diagnosis  Applying text mining to extract Framingham symptoms can help trigger early intervention 3,168 patients eventually all diagnosed with HF Vhavakrishnan R, Steinhubl SR, Sun J, et al. Potential impact of predictive models for early detection of heart failure on the initiation of evidence-based therapies. J Am Coll Cardiol. 2012;59(13s1):E949-E949. No symptoms Framingham symptoms Clinical diagnosis Opportunity for early intervention
  • 57.
  • 58.
    58 KNOWLEDGE+DATA FEATURE SELECTION Jimeng Sun,Jianying Hu, Dijun Luo, Marianthi Markatou, Fei Wang, Shahram Ebadollahi, Steven E. Steinhubl, Zahra Daar, Walter F. Stewart. Combining Knowledge and Data Driven Insights for Identifying Risk Factors using Electronic Health Records. AMIA (2012).
  • 59.
    59 Combining Knowledge- andData-driven Risk Factors Knowledge Data Combination Risk factor augmentation Risk factor augmentation Combined risk factors Combined risk factors Knowledge base Knowledge base Risk factor gathering Risk factor gathering Knowledge risk factors Knowledge risk factors Clinical dataClinical data Data processing Data processing Potential risk factors Potential risk factors Target condition Target condition Risk factor augmentation Risk factor augmentation
  • 60.
    60 Risk Factor Augmentation Model Accuracy: – The selected risk factors are highly predictive of the target condition – Sparse feature selection through L1 regularization  Minimal Correlations: – Between data driven risk factors and knowledge driven risk factors – Among the data driven risk factors Model error Correlation among data-driven features Correlation between data- and knowledge- driven features Sparse Penalty Dijun Luo, Fei Wang, Jimeng Sun, Marianthi Markatou, Jianying Hu,Shahram Ebadollahi, SOR: Scalable Orthogonal Regression for Low-Redundancy Feature Selection and its Healthcare Applications. SDM’12
  • 61.
    61 Prediction Results usingSelected Features  AUC significantly improves as complementary data driven risk factors are added into existing knowledge based risk factors.  A significant AUC increase occurs when we add first 50 data driven features +200+150 +100 +50 all knowledge features+diabetes +Hypertension CAD 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0 100 200 300 400 500 600 AUC Number of features Jimeng Sun, Jianying Hu, Dijun Luo, Marianthi Markatou, Fei Wang, Shahram Ebadollahi, Steven E. Steinhubl, Zahra Daar, Walter F. Stewart. Combining Knowledge and Data Driven Insights for Identifying Risk Factors using Electronic Health Records. AMIA (2012).
  • 62.
    62 Clinical Validation ofData-driven Features  9 out of 10 are considered relevant to HF  The data driven features are complementary to the existing knowledge-driven features
  • 63.
  • 64.
  • 65.
    65 Anatomy of ClinicalPredictive Model  Prediction Models –Continuous outcome: Regression –Categorical outcome: Classification • Logistic regression –Survival outcome • Cox Proportional Hazard Regression –Patient Similarity  Case study: Heart failure onset prediction
  • 66.
  • 67.
    67 Intuition of PatientSimilarity Patient Doctor Similarity search
  • 68.
    68 Intuition of PatientSimilarity Patient Doctor
  • 69.
    69 Summary on PatientSimilarity  Patient similarity learns a customized distance metric for a specific clinical context  Extension 1: Composite distance integration (Comdi) [SDM’11a] – How to jointly learn a distance by multiple parties without data sharing?  Extension 2: Interactive metric update (iMet) [SDM’11b] – How to interactively update an existing distance measure? 1. Jimeng Sun, Fei Wang, Jianying Hu, Shahram Edabollahi: Supervised patient similarity measure of heterogeneous patient records. SIGKDD Explorations 14(1): 16-24 (2012) 2. Fei Wang, Jimeng Sun, Shahram Ebadollahi: Integrating Distance Metrics Learned from Multiple Experts and its Application in Inter-Patient Similarity Assessment. SDM 2011: 59-70 56 3. Fei Wang, Jimeng Sun, Jianying Hu, Shahram Ebadollahi: iMet: Interactive Metric Learning in Healthcare Applications. SDM 2011: 944-955
  • 70.
  • 71.
    71 Motivations for EarlyDetection of Heart Failure  Heart failure (HF) is a complex disease  Huge Societal Burden  For payers – Reduce cost and hospitalization – Improve the existing clinical guidance of HF prevention  For providers – Slow or potentially reverse disease progress – Improve quality of life, reduce mortality
  • 72.
    72 Predictive Modeling StudyDesign  Goal: Classify HF cases against control patients  Population – 50,625 Patients (Geisinger Clinic PCPs) – Cases: 4,644 case patients – Controls 45,981 matched on age, gender and clinic Cases Controls
  • 73.
    73 Predictive Modeling Setup We define – Diagnosis date and index date – Prediction and observation windows  Features are constructed from the observation window and predict HF onset after the prediction window Index date Diagnosis date Observation Window Prediction Window
  • 74.
    74 Features  We constructover 20K features of different types  Through feature selection and generalization, we result in the following predictive features Feature type Cardinality Predictive Features DIAGNOSIS 17,322 Diabetes, CHD, hypertensions, valvular disease, left ventricular hypertrophy, angina, atrial fibrillation, MI, COPD Demographics 11 Age, race, gender, smoking status Framingham 15 rales, cardiomegaly, acute pulmonary edema, HJReflex, ankle edema, nocturnal cough, DOExertion, hepatomegaly, pleural effusion Lab 1,264 eGFR, LVEF, albumin, glucose, cholesterol, creatinine, cardiomegaly, heart rate, hemoglobin Medication 3,922 antihypertensive, lipid-lowering, CCB, ACEI, ARB, beta blocker, diuretic, digitalis, antiarrhythmic Vital 6 blood pressure and heart rate
  • 75.
    75 Prediction Performance onDifferent Prediction Windows  Setting: observation window = 720 days, classifiers = random forest, evaluation mechanism = 10-fold cross-validation for 10 times  Observation:  AUC slowly decreases as the prediction window increases 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 0 90 180 270 360 450 540 630 720 AUC Prediction Window (days) HF Onset (ObservationWindow=720, PredictionWindow=variable)
  • 76.
    76 Prediction Performance onDifferent Observation Windows  Setting: prediction window= 180 days, classifiers= random forest, evaluation mechanism =10-fold cross-validation  Observation:  AUC increases as the observation window increases. i.e., more data for a longer period of time will lead to better performance of the predictive model  Combined features performed the best at observation window = 720 days 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 30 90 180 270 360 450 540 630 720 810 900 AUC Observation Window (days) HF Onset (ObservationWindow=variable, PredictionWindow=180)
  • 77.
  • 78.
  • 79.
    79 Unstructured Clinical Data DatasetLink Description i2b2 Informatics for Integrating Biology & the Bedside https://www.i2b2.org/NLP/Dat aSets/Main.php Clinical notes used for clinical NLP challenges • 2006 Deidentification and Smoking Challenge • 2008 Obesity Challenge • 2009 Medication Challenge • 2010 Relations Challenge • 2011 Co-reference Challenge Computational Medicine center http://computationalmedicine. org/challenge/previous Classifying Clinical Free Text Using Natural Language Processing
  • 80.
    80 Structured EHR Dataset LinkDescription Texas Hospital Inpatient Discharge http://www.dshs.state.tx.us/thcic/ho spitals/Inpatientpudf.shtm Patient: hospital location, admission type/source, claims, admit day, age, icd9 codes + surgical codes Framingham Health Care Data Set http://www.framinghamheartstudy.o rg/share/index.html Genetic dataset for cardiovascular disease Medicare Basic Stand Alone Claim Public Use Files http://resdac.advantagelabs.com/c ms-data/files/bsa-puf Inpatient, skilled nursing facility, outpatient, home health agency, hospice, carrier, durable medical equipment, prescription drug event, and chronic conditions on an aggregate level VHA Medical SAS Datasets http://www.virec.research.va.gov/M edSAS/Overview.htm Patient care encounters primarily for Veterans: inpatient/outpatient data from VHA facilities Nationwide Inpatient Sample http://www.hcup- us.ahrq.gov/nisoverview.jsp Discharge data from 1051 hospitals in 45 states with diagnosis, procedures, status, demographics, cost, length of stay CA Patient Discharge Data http://www.oshpd.ca.gov/HID/Produ cts/PatDischargeData/PublicDataS et/index.html Discharge data for licensed general acute hospital in CA with demographic, diagnostic and treatment information, disposition, total charges MIMIC II Clinical Database http://mimic.physionet.org/database .html ICU data including demographics, diagnosis, clinical measurements, lab results, interventions, notes Thanks to Prof. Joydeep Ghosh from UT Austin for providing this information
  • 81.
    81 Software  MetaMap mapsbiomedical text to UMLS metathesaurus – Developed by NLM for parsing medical article not clinical notes – http://metamap.nlm.nih.gov/  cTAKES: clinical Text Analysis and Knowledge Extraction System – Using Unstructured Information Management Architecture (UIMA) framework and OpenNLP toolkit – http://ctakes.apache.org/
  • 82.
    82 Organization of thisTutorial  Introduction  Motivating Examples  Sources and Techniques for Big Data in Healthcare – Structured EHR Data – Unstructured Clinical Notes – Medical Imaging Data – Genetic Data – Other Data (Epidemiology & Behavioral)  Final Thoughts and Conclusion
  • 83.
  • 84.
    84 Image Data isBig !!!  By 2015, the average hospital will have two-thirds of a petabyte (665 terabytes) of patient data, 80% of which will be unstructured image data like CT scans and X-rays.  Medical Imaging archives are increasing by 20%-40%  PACS (Picture Archival & Communication Systems) system is used for storage and retrieval of the images. Image Source: http://medcitynews.com/2013/03/the-body-in-bytes-medical-images-as-a-source-of-healthcare-big-data-infographic/
  • 85.
    85 Popular Imaging Modalitiesin Healthcare Domain Computed Tomography (CT)  The main challenge with the image data is that it is not only huge, but is also high-dimensional and complex.  Extraction of the important and relevant features is a daunting task.  Many research works applied image features to extract the most relevant images for a given query. Positron Emission Tomography (PET) Magnetic Resonance Imaging (MRI) Image Source: Wikipedia
  • 86.
    86 Medical Image RetrievalSystem Biomedical Image Database Testing Phase Feature Extraction Algorithms for learning or similarity computations Final Trained Models Query Image Retrieval System Query Results Performance Evaluation (Precision-Recall) Recall Precision Training Phase
  • 87.
    87 Content-based Image Retrieval Two components – Image features/descriptors - bridging the gap between the visual content and its numerical representation. – These representations are designed to encode color and texture properties of the image, the spatial layout of objects, and various geometric shape characteristics of perceptually coherent structures. – Assessment of similarities between image features based on mathematical analyses, which compare descriptors across different images. – Vector affinity measures such as Euclidean distance, Mahalanobis distance, KL divergence, Earth Mover’s distance are amongst the widely used ones.
  • 88.
    88 Medical Image Features Photo-metricfeatures exploit color and texture cues and they are derived directly from raw pixel intensities. Geometric features: cues such as edges, contours, joints, polylines, and polygonal regions. Akgül, Ceyhun Burak, et al. "Content-based image retrieval in radiology: current status and future directions." Journal of Digital Imaging 24.2 (2011): 208-222. Müller, Henning, et al. "A review of content-based image retrieval systems in medical applications-clinical benefits and future directions." International journal of medical informatics 73.1 (2004): 1-24. • A suitable shape representation should be extracted from the pixel intensity information by region-of interest detection, segmentation, and grouping. Due to these difficulties, geometric features are not widely used.
  • 89.
    89 Image CLEF Data ImageCLEF aims to provide an evaluation forum for the cross– language annotation and retrieval of images (launched in 2003)  Statistics of this database : – With more than 300,000 (in .JPEG format), the total size of the database > 300 GB – contains PET, CT, MRI, and Ultrasound images  Three Tasks – Modality classification – Image–based retrieval – Case–based retrieval Medical Image Database available at http://www.imageclef.org/2013/medical
  • 90.
    90 Modality Classification Task Modalityis one of the most important filters that clinicians would like to be able to limit their search by.
  • 91.
    91 Image based andCase-based Querying  Image-based retrieval : This is the classic medical retrieval task. Similar to Query by Image Example. Given the query image, find the most similar images.  Case-based retrieval: This is a more complex task; is closer to the clinical workflow. A case description, with patient demographics, limited symptoms and test results including imaging studies, is provided (but not the final diagnosis). The goal is to retrieve cases including images that might best suit the provided case description.
  • 92.
    92 Challenges with ImageData  Extracting informative features.  Selection of relevant features.  Sparse methods* and dimensionality reducing techniques  Integration of Image data with other data available  Early Fusion  Vector-based Integration  Intermediate Fusion  Multiple Kernel Learning  Late Fusion  Ensembling results from individual modalities *Jieping Ye’s SDM 2010 Tutorial on Sparse methods http://www.public.asu.edu/~jye02/Tutorial/Sparse-SDM10.pdf
  • 93.
    93 Publicly Available MedicalImage Repositories Image database Name Moda lities No. Of patients No. Of Images Size Of Data Notes/Applications Download Link Cancer Imaging Archive Database CT DX CR 1010 244,527 241 GB Lesion Detection and classification, Accelerated Diagnostic Image Decision, Quantitative image assessment of drug response https://public.cancerimagingarchive.net/ ncia/dataBasketDisplay.jsf Digital Mammog raphy database DX 2620 9,428 211 GB Research in Development of Computer Algorithm to aid in screening http://marathon.csee.usf.edu/Mammogr aphy/Database.html Public Lung Image Database CT 119 28,227 28 GB Identifying Lung Cancer by Screening Images https://eddie.via.cornell.edu/crpf.html Image CLEF Database PET CT MRI US unknown 306,549 316 GB Modality Classification , Visual Image Annotation , Scientific Multimedia Data Management http://www.imageclef.org/2013/medical MS Lesion Segment ation MRI 41 145 36 GB Develop and Compare 3D MS Lesion Segmentation Techniques http://www.ia.unc.edu/MSseg/download .php ADNI Database MRI PET 2851 67,871 16GB Define the progression of Alzheimer’s disease http://adni.loni.ucla.edu/data- samples/acscess-data/
  • 94.
  • 95.
    95 Genetic Data  Thehuman genome is made up of DNA which consists of four different chemical building blocks (called bases and abbreviated A, T, C, and G).  It contains 3 billion pairs of bases and the particular order of As, Ts, Cs, and Gs is extremely important.  Size of a single human genome is about 3GB.  Thanks to the Human Genome Project (1990-2003) – The goal was to determine the complete sequence of the 3 billion DNA subunits (bases). – The total cost was around $3 billion.
  • 96.
    96 Genetic Data  Thewhole genome sequencing data is currently being annotated and not many analytics have been applied so far since the data is relatively new.  Several publicly available genome repositories. http://aws.amazon.com/1000genomes/  It costs around $5000 to get a complete genome. It is still in the research phase. Heavily used in the cancer biology.  In this tutorial, we will focus on Genome-Wide Association Studies (GWAS). – It is more relevant to healthcare practice. Some clinical trials have already started using GWAS. – Most of the computing literature (in terms of analytics) is available for the GWAS. It is still in rudimentary stage for whole genome sequences.
  • 97.
    97 Genome-Wide Association Studies(GWAS)  Genome-wide association studies (GWAS) are used to identify common genetic factors that influence health and disease.  These studies normally compare the DNA of two groups of participants: people with the disease (cases) and similar people without (controls). (One million Loci)  Single nucleotide polymorphisms (SNPs) are DNA sequence variations that occur when a single nucleotide (A,T,C,or G) in the genome sequence differs between individuals.  SNPs occur every 100 to 300 bases along the 3-billion-base human genome.
  • 98.
    98 Epistasis Modeling  Forsimple Mendelian diseases, single SNPs can explain phenotype very well.  The complex relationship between genotype and phenotype is inadequately described by marginal effects of individual SNPs.  Increasing empirical evidence suggests that interactions among loci contribute broadly to complex traits.  The difficulty in the problem of detecting SNP pair interactions is the heavy computational burden. – To detect pairwise interactions from 500,000 SNPs genotyped in thousands of samples, a total of 1.25 X 10 statistical tests are needed.
  • 99.
    99 Epistasis Detection Methods Exhaustive – Enumerates all K-locus interactions among SNPs. – Efficient implementations mostly aiming at reducing computations by eliminating unnecessary calculations.  Non-Exhaustive – Stochastic: randomized search. Performance lowers when the # SNPs increase. – Heuristic: greedy methods that do not guarantee optimal solution. Shang, Junliang, et al. "Performance analysis of novel methods for detecting epistasis." BMC bioinformatics 12.1 (2011): 475.
  • 100.
    100 Sparse Methods forSNP data analysis  Successful identification of SNPs strongly predictive of disease promises a better understanding of the biological mechanisms underlying the disease.  Sparse linear methods have been used to fit the genotype data and obtain a selected set of SNPs.  Minimizing the squared loss function (L) of N individuals and p variables (SNPs) is used for linear regression and is defined as where xi ∈ ℝp are inputs for the ith sample, y ∈ ℝN is the N vector of outputs, β0 ∈ ℝ is the intercept, β ∈ ℝp is a p-vector of model weights, and λ is user penalty.  Efficient implementations that scale to genome-wide data are available.  SparSNP package http://bioinformatics.research.nicta.com.au/software/sparsnp/ Wu, Tong Tong, et al. "Genome-wide association analysis by lasso penalized logistic regression." Bioinformatics 25.6 (2009): 714-721.     N i p j j T ii xyL 1 1 2 00 )( 2 1 ),( 
  • 101.
    101 Public Resources forGenetic (SNP) Data  The Wellcome Trust Case Control Consortium (WTCCC) is a group of 50 research groups across the UK which was established in 2005.  Available at http://www.wtccc.org.uk/  Seven different diseases: bipolar disorder (1868 individuals), coronary heart disease (1926 individuals), Crohn's disease (1748 individuals), hypertension (1952 individuals), rheumatoid arthritis (1860 individuals), type I diabetes (1963 individuals) or type II diabetes (1924 individuals).  Around 3,000 healthy controls common for these disorders. The individuals were genotyped using Affymetrix chip and obtained approximately 500K SNPs.  The database of Genotypes and Phenotypes (dbGaP) maintained by National Center of Biotechnology Information (NCBT) at NIH.  Available at http://www.ncbi.nlm.nih.gov/gap
  • 102.
  • 103.
    103 Epidemiology Data  TheSurveillance Epidemiology and End Results Program (SEER) at NIH.  Publishes cancer incidence and survival data from population-based cancer registries covering approximately 28% of the population of the US.  Collected over the past 40 years (starting from January 1973 until now).  Contains a total of 7.7M cases and >350,000 cases are added each year.  Collect data on patient demographics, tumor site, tumor morphology and stage at diagnosis, first course of treatment, and follow-up for vital status. Usage:  Widely used for understanding disparities related to race, age, and gender.  Can be used to overlay information with other sources of data (such as water/air pollution, climate, socio-economic) to identify any correlations.  Can not be used for predictive analysis, but mostly used for studying trends. SEER database is available at http://seer.cancer.gov/
  • 104.
    104 Social Media canSense Public Health !! Chunara, Rumi, Jason R. Andrews, and John S. Brownstein. "Social and news media enable estimation of epidemiological patterns early in the 2010 Haitian cholera outbreak." American Journal of Tropical Medicine and Hygiene 86.1 (2012): 39. During infectious disease outbreaks, data collected through health institutions and official reporting structures may not be available for weeks, hindering early epidemiologic assessment. Social media can get it in near real-time. Dugas, Andrea Freyer, et al. "Google Flu Trends: correlation with emergency department influenza rates and crowding metrics." Clinical infectious diseases 54.4 (2012): 463-469. Google Flu Trends correlated with Influenza outbreak Twitter messaging correlated with cholera outbreak
  • 105.
    105 Social Networks forPatients  PatientsLikeMe1 is a patient network is an online data sharing platform started in 2006; now has more than 200,000 patients and is tracking 1,500 diseases. 1 http://www.patientslikeme.com/ OBJECTIVE: “Given my status, what is the best outcome I can hope to achieve, and how do I get there?”  People connect with others who have the same disease or condition, track and share their own experiences, see what treatments have helped other patients like them, gain insights and identify any patterns.  Patient provides the data on their conditions, treatment history, side effects, hospitalizations, symptoms, disease-specific functional scores, weight, mood, quality of life and more on an ongoing basis.  Gaining access to the patients for future clinical trials.
  • 106.
    106 Home Monitoring andSensing Technologies  Advancements in sensing technology are critical for developing effective and efficient home-monitoring systems • Sensing devices can provide several types of data in real-time. • Activity Recognition using Cell Phone Accelerometers Kwapisz, Jennifer R., Gary M. Weiss, and Samuel A. Moore. "Activity recognition using cell phone accelerometers." ACM SIGKDD Explorations Newsletter 12.2 (2011): 74-82. Rashidi, Parisa, et al. "Discovering activities to recognize and track in a smart environment." Knowledge and Data Engineering, IEEE Transactions on 23.4 (2011): 527-539.
  • 107.
    107 Public Health andBehavior Data Repositories Dataset Link Description Behavioral Risk Factor Surveillance System (BRFSS) http://www.cdc.gov/brfss/technical_ infodata/index.htm Healthcare survey data: smoking, alcohol, lifestyle (diet, exercise), major diseases (diabetes, cancer), mental illness Ohio Hospital Inpatient/Outpatient Data http://publicapps.odh.ohio.gov/pwh /PWHMain.aspx?q=021813114232 Hospital: number of discharges, transfers, length of stay, admissions, transfers, number of patients with specific procedure codes US Mortality Data http://www.cdc.gov/nchs/data_acce ss/cmf.htm Mortality information on county- level Human Mortality Database http://www.mortality.org/ Birth, death, population size by country Utah Public Health Database http://ibis.health.utah.gov/query Summary statistics for mortality, charges, discharges, length of stay on a county-level basis Dartmouth Atlas of Health Care http://www.dartmouthatlas.org/tools /downloads.aspx Post discharge events, chronically ill care, surgical discharge rate Thanks to Prof. Joydeep Ghosh from UT Austin for providing this information.
  • 108.
  • 109.
    109 Final Thoughts Big datacould save the health care industry up to $450 billion, but other things are important too.  Right living: Patients should take more active steps to improve their health.  Right care: Developing a coordinated approach to care in which all caregivers have access to the same information.  Right provider: Any professionals who treat patients must have strong performance records and be capable of achieving the best outcomes.  Right value: Improving value while simultaneously improving care quality.  Right innovation: Identifying new approaches to health-care delivery. “Stakeholders will only benefit from big data if they take a more holistic, patient-centered approach to value, one that focuses equally on health-care spending and treatment outcomes,” McKinsey report available at: http://www.mckinsey.com/insights/health_systems/the_big-data_revolution_in_us_health_care
  • 110.
    110 Conclusion  Big dataanalytics is a promising right direction which is in its infancy for the healthcare domain.  Healthcare is a data-rich domain. As more and more data is being collected, there will be increasing demand for big data analytics.  Unraveling the “Big Data” related complexities can provide many insights about making the right decisions at the right time for the patients.  Efficiently utilizing the colossal healthcare data repositories can yield some immediate returns in terms of patient outcomes and lowering care costs.  Data with more complexities keep evolving in healthcare thus leading to more opportunities for big data analytics.
  • 111.
    111 Acknowledgements • Funding Sources –National Science foundation – National Institutes of Health – Susan G. Komen for the Cure – Delphinus Medical Technologies – IBM Research
  • 112.
    112 Thank You Questions andComments Feel free to email questions or suggestions to [email protected] [email protected]