OHDSI Home | Forums | Wiki | Github

[Patient-Level Prediction Workgroup] Please Introduce Yourself!

Researchers
#1

Dear all,

As mentioned in the first PLP TC i would like to ask all of you to introduce yourselves (if you do not have that information on your OHDSI profile) and inform the rest of the team about your experience and interest in this topic.

If we find common interests we could create smaller teams that focus on specific aspects, e.g. the use of temporal information, feature engineering etc.

Let me kick this off:

You can find information on my background on my profile. I am very happy to have the opportunity to lead this interesting workgroup together with the great Jenna Reps :slight_smile: My interest are on all the aspects of patient-level prediction, with a focus on the use of temporal information and incorporation of expert knowledge. I will be working on this intensively the next years via a project between Erasmus MC and Janssen (Patrick Ryan). Let me say that i will predict this will be a nice year ahead!

#2

Hello, I’m Jenna, my background over the past 5 years has been applying various machine learning methods to solve healthcare related issues, with my phd focusing on predicting adverse drug reactions. I’m keen to contribute by working on incorporating new feature engineering/selection methods and high performance classifiers into the patient level prediction library. I’m also happy to help out with any R/python code that is needed. I hope to see some nice personalised models being trained by the team over the next year :slight_smile:

#3

Hello everyone! I’m Narges Razavian. I’m a postdoc at Clinical Machine Learning lab lead by Dr. David Sontag. My background is machine learning for bioinformatics and medical informatics.

Our most recent publication is a large study, where we built 42,000 health features (including temporal features) from administrative claims data, to learn a model for predicting Type 2 diabetes onset. We trained and independently tested our model on data from an initial cohort of 4.1 million individuals over 8 years.

You can read it here:
Population-Level Prediction of Type 2 Diabetes From Claims Data and Analysis of Risk Factors
Narges Razavian, Saul Blecker, Ann Marie Schmidt, Aaron Smith-McLallen, Somesh Nigam, and David Sontag, Big Data. January 2016.

We are converting our full cohort into common data format developed by OHDSI, and plan to release our models (not just for diabetes, for all diseases) compatible with this standard data format, so everyone could plug them into their data and get predictions. I will present this paper on the meeting of Jan 27th, 2016, and look forward to collaborating on the next steps!

We have also built a more advanced machine learning model for detecting (3 months in advance) multiple (270) diseases, using temporal convolution networks and lab values of the individual in the past 3 years. I can discuss that one further into the future! This work is still under review: http://arxiv.org/abs/1511.07938

#4

Hi, I’m Rich Boyce
(http://www.ohdsi.org/who-we-are/collaborators/richard-d-boyce-2/). I am
interested in developing an effective informatics interventions that
prevent harm to nursing home residents from while avoiding known issues
such as alert fatigue. We are building patient-level prediction models
for falls and other outcomes using nursing home data and so I consider
myself an end user of what the working group is developing. Also, I am
very interested how to help build PLP into an overall framework for
helping health systems design and implement effective clinical decision
support. I think I am my team can apply the methods and help build user
tools that fit in the OHDSI framework for meeting the goals I just
mentioned.

best,
-R

#5

Hi, Martijn Schuemie here. I’m the one mostly responsible for the current PatientLevelPrediction R package. I work at J&J, where we have several observational databases in the OMOP CDM.

I’m interested in many things, but one thing on top of my list is scaling patient level prediction to a large number of outcomes. I think that even though existing methods canbe improved, there could be large value in applying them to real problems and see if they can be used to help people.

#6

Hi. I’m Kenney Ng. I am a research staff member in the Health Informatics Department and manager of the Healthcare Analytics Research Group at IBM Thomas J. Watson Research Center. My team focusses on the research, development and application of data mining and machine learning techniques to analyze, model and derive insights from real world healthcare data. Areas of research include similarity analytics, predictive modelling, interactive visual analytics, disease modelling, translational medicine and scalable analytic platforms. In the space of patient-level prediction, I am interested in developing approaches that can build more precise/personalized predictive models from appropriate subsets of the data. I hope to be able to contribute ideas and code to the community.

#7

Hi everyone, I’m Jamie, a new member of the Epidemiology Analytics team at Janssen R&D. My background is in epidemiology and applied statistics and I’ve spent some time at universities in various capacities working on comparative effectiveness studies, methods for causal inference, and some health services research. More recently, I’ve become interested in data mining and predictive analytics and spent the last year and a half studying methods for supervised and unsupervised modeling, feature selection, and model performance evaluation. I’m very interested in how high performance models can be standardized and applied across diverse outcomes and problems, much like what’s being done here. I look forward to collaborating with the OHDSI community and contributing wherever I can.

#8

Team,

I like to introduce a new member to our team: Prof. Joydeep Ghosh from the Department of Electrical & Computer Engineering Univ of Texas, Austin. @joydeep

You can find more information about his research group here: http://ideal.ece.utexas.edu/ghosh/

Peter

#9

Hello. My name is Johan van der Lei. I work at Erasmus Medical Center (Dept. of Medical Informatics) with an interest in decisions support and patient-level prediction. Here in the Department, we do have a data base derived from the records of general practitioners (the so-called IPCI data base). The data base contains detailed medical data (including free text notes) from approx 2 million Dutch people. One of our objectives is to use that data base also in the context of patient-level prediction. As a group, we have a history in trying to combine different European data bases in order to conduct studies in different countries.

1 Like
#10

Dear all,

I am Associate Prof Dani Prieto-Alhambra, from NDORMS, University of Oxford. I am an academic GP and NIHR Clinician Scientist with expertise and interest in patient prediction using routinely collected data (EMR and registry data). This is my profile web and my group’s website for your info.

I am also scientific coordinator for the SIDIAP Database (www.sidiap.org), a Spanish emr database covering >80% of the population of Catalonia. SIDIAP will be in the OMOP CDM this year, so looking forward to working with you all on that as well!

Peter has kindly let me know about this group, and I will be delighted to contribute where possible.

All best everyone!

#11

Hello, everyone. I am Zixiao Shen, which is a PhD student in the School of Computer Science, University of Nottingham, UK. You can also call me Zeal Shen. My research interest is in the medical data mining. By using the machine learning and data mining techniques on the medical database such as CPRD and so on, I am aimed to discover some interesting ideas and knowledge.

Dr. Reps has kindly recommended me to the project. And I am looking forward to contributing wherever I can and communicating with others.

#12

Hi everyone, I’m S’busiso Mkhondwane. I’m a senior consultant at Avanade South Africa. My background is in business intelligence and have interest in R, machine learning and advanced analytics. I used to be an intern at SAP Research focusing on the public health. I would like to use this platform to further my studies with a focus on advanced analytics within the public health sector. Looking forward to learning and gaining knowledge from the network and providing/sharing my experience as well.

#13

I am an electronics engineer by training and my research has primarily focused on signal processing and machine learning for early detection of deterioration in patients. This work involves computational modelling of patient’s vital signs using data fusion and novelty detection methods. This research is now moving towards patient-specific models. Currently I am working on 1) developing a classification model to enable clinicians to diagnose and distinguish between different types of vasculitis, 2) machine learning alternative for propensity score matching, 3) prediction of chronic pain after knee surgery using post-operative risk factors. Please find my details here.

#14

Hi My name is Feifan Liu, an assistant professor at University of Massachusetts Medical School, with a background of natural language processing and machine learning. We are doing research on detecting adverse drug events on EHR narrative data, and like to explore predictive modeling on patient level on a large scale analysis integrating multiple disparate data sources. I learned a lot at 2017 face to face meeting and with Jenna and Peter’s help I was able to add another classifier in the platform. Hope to do more collaboration within this group in the future.

#15

Hi Feifan Liu,

It was very nice meeting you at the face-to-face. If you send me a personal message on the forum with your email address, affiliation etc I will add you on the wiki page and will contact you so we can find ways to collaborate.

Peter

#16

Hi Everyone,

I am Xiaoyong Pan, a postdoc at Department of medical informatics, Erasmus MC. i have a background of bioinformatics, machine learning and deep learning. I worked on deep learning for biological data in genomic sequences during my PhD study at Copenhagen University. Now i am working on applying deep learning, e.g. convolutional neural network and recurrent neural network, on electronic health data, especially temporal data. I am looking forward to collaborating on this topic.

1 Like
#17

@xypan1232
Hi, Xiaoyong.

I’m Seng Chan You. I worked as an internist in the hospital, and now I am studying medical informatics in Ajou university, Korea.
I and some members in Ajou university are also interested in the same topic. We’re trying to adopt some existing deep learning algorithm on CDM, now.
We’re happy to collaborate with you about this.

Cheers,
Chan

#18

Hello everyone.

I heard about the OHDSI initiative in a talk given by Professor David Madigan in the Bayesian Statistics Symposium 2018, University of Arizona.

I am fourth year PhD in the Management Information Systems Department of University of Arizona and you can find out more about me and my research interests in my webpage: http://karanalytics.com/.

I am interested in modeling disease progression in the context of disease comorbidity. That is, how what are simultaneous occurrences of multiple diseases existing in a patient encounter related to the onset of future diseases.

I am very new to the OHDSI community. I do not have a dataset, and I am open to collaborating with others.

#19

Hi All,
My name is Igor Korolev, and I’m an incoming Medical Informatics Fellow at the Boston VA Healthcare System. By training, I’m a physician-neuroscientist, with a passion for clinical research at the interface of healthcare and technology. In addition to a long-standing interest and experience in analyzing large biomedical data sets relating to aging, cognition, and neuropsychiatric disorders (behavioral, neuroimaging, EEG, etc), including Alzheimer’s disease, I’m very interested in real-world health data analysis (i.e. EHR and wearables). Specifically, I’m interested in using machine learning, informatics, and other data science approaches to advance precision medicine and population health as well as to develop clinical decision support tools. I look forward to contributing and collaborating with the OHDSI community!

Kind regards,
Igor

t