2 posts tagged “predictive modeling”

PatientsLikeMe Researchers Score a Hat Trick

Posted July 11th, 2012 by

The term “hat trick” originated in 1858 after English cricketer HH Stephenson successfully bowled out three batsmen with consecutive bowls and he was presented with a hat to commemorate his feat. In June, PatientsLikeMe’s research team scored a hat trick of our own by publishing three new studies in scientific journals in just four days, bringing our total number of published studies to 27.

An Image from the Third Published Study in Our June Hat Trick:  "Mining Social Network Data for Biomedical Research"

As employees at a start-up company, we all wear many hats – literally, as you can see from the photos below, and figuratively, in terms of our responsibilities as scientists, product developers and business people across different disease areas. For instance, in this trio of papers, we address research issues in multiple sclerosis (MS), ALS and mood disorders. Click on the study titles below to read the full papers and a big thank you to all of our patients for sharing your voices and making this research possible.

Monday, June 18th – First PatientsLikeMe study published.

The multiple sclerosis rating scale, revised (MSRS-R): Development, refinement, and psychometric validation using an online community

Members of our MS community will be familiar with the MS Rating Scale (MSRS), which is their primary outcome measure. We developed it a few years ago to address an unmet need for a brief, easy-to-use rating tool that covered more areas than simply walking. In collaboration with a neurologist, we sketched out our first version of the MSRS, which has now been used over 90,000 times by our 28,000+ MS members to share your progress, track your relapses and disability, and gain insight into how your treatments are working.

PatientsLikeMe Research Scientist Dr. Tim Vaughan

Published in the Journal of Medical Internet Research, this new study describes work to improve the MSRS to a revised version (the MSRS-R) and establish that the instrument measures what it is supposed to measure (reliability), that it correlates well with other instruments (concurrent validity) and that it allows differences between groups of patients (sensitivity to change). This process of “validating” a patient-reported outcome (PRO) is an important step in increasing the value of the data that we produce for researchers. Studies are underway right now to continue improving the instrument and compare it to doctors’ ratings of patient disability.

As with all outcome measures we develop at PatientsLikeMe, we have licensed the MSRS-R for anyone to use freely in their own research studies. This work also serves as a foundation for our resident predictive modeler and particle physicist Dr. Tim Vaughan to begin work on predicting the course of an individual patient’s disease using your MSRS scores!

Tuesday, June 19th – Second PatientsLikeMe study published.

E-mental health: A medium reaches maturity

PatientsLikeMe Research & Development Director Dr. Paul Wicks

The Internet has transformed many aspects of healthcare in the past decade, and to open a special issue on “E-Mental Health” in the Journal of Mental Health, our R&D Director Dr. Paul Wicks was commissioned to write a special editorial.

Available by clicking the link above, the paper describes the progress of online systems for people with mental health issues, from government-provided resources (such as this UK NHS Choices site about self-harm) to commercial, computerized cognitive-behavioral therapy programs like “Beating the Blues” and collaborations between different sectors of the health system. For instance, in the UK if you Google “suicide” there is a special message from the Samaritans mental health support service right at the top of the page.

At PatientsLikeMe, our vibrant mood community has been active since 2008, and our published research has shown that it provides improved outcomes to many of our members.

Thursday, June 21st – Third PatientsLikeMe study published.  Hat trick completed.

Mining Online Social Network Data for Biomedical Research: A Comparison of Clinicians’ and Patients’ Perceptions About Amyotrophic Lateral Sclerosis Treatments

As you probably know, the story of PatientsLikeMe starts with ALS and trying to find treatments that can improve the symptoms, including anxiety, stiffness, or constipation. Back in 2003, ALS nurse Dallas Forshew and Dr. Mark Bromberg published a small study describing massive variation in the way ALS doctors from 39 specialist centers treat the symptoms of ALS. This data was also described in our 2010 TEDx Berkshires talk about the value of crowd-sourced data.

PatientsLikeMe Research Assistant/Software Engineer Shivani Bhargava

In this new study published in the Journal of Medical Internet Research, we collaborated with researchers at the University of Utah to compare the passively collected ALS patient data from our system with what the clinicians said. Although they agreed in most areas, there were split opinions too, particularly when it came to the perceived level of efficacy that these treatments had. Click the link above to read more.

This paper is also notable for being our own Shivani Bhargava’s very first scientific publication! Shivani started with us as an intern, then became a research assistant, and has recently made a career change to start studying as a software engineer. A true renaissance woman!


Charting the course of PLS and PMA

Posted August 11th, 2009 by

Primary Lateral Sclerosis (PLS) and Progressive Muscular Atrophy (PMA) are two rare variants of the disease ALS. Normally, ALS affects the upper motor neurons in the brain and spinal cord, as well as the lower motor neurons that send signals from the spinal cord directly to muscles. PLS and PMA are different because PLS only affects the upper motor neurons, and PMA only affects the lower motor neurons. It’s an important distinction for patients to be told about because the prognosis is less severe in these conditions.  On average, survival in ALS is typically reported to be 2-5 years, whereas for patients with PMA it’s more like 5-10 years and for PLS it’s even longer (often several decades).

ALS itself is a rare condition, affecting some 30,000 people in the United States at any one time.  PLS and PMA each represent approximately 5% of the overall ALS community, so there’s approximately 1,500 patients with each condition in the U.S at any given time.  In April 2008, PatientsLikeMe added the ability for members of our ALS community to change their diagnosis to these rare conditions.  To date, we now have 182 patients with PLS and 270 with PMA. This is truly exciting because even the largest studies in the literature have only examined 40 or so PLS patients and a similar number of PMA patients. One of our most useful features on our site for people with ALS is the percentile curves, which we display as a backdrop on their profiles to put each individual’s rate of progression into context. However, as you can see in the figure below, when you compare the progression curves of ALS patients on our site with those of a typical PLS patient, the PLS patient progression deviates significantly from the ALS curves.

pls-patient-on-als-curves_ls2

With so many PLS and PMA patients sharing such valuable information about their disease on PatientsLikeMe, we had enough information to generate a new set of percentile curves for each of those communities.  To do this, we used self-report ALSFRS-R (ALS functional rating score – revised) data from 104 PLS patients and 59 PMA patients that met our criteria for data quality. We have good data for the first 4-5 years of disease after onset, and after that point we rely on linear extrapolation to make the plots.  Here we see the value of openness in action.  When you see the potential value in contributing your data, it drives a virtuous cycle: the more data you enter, the more value you get, so you enter more data!

pma_profile1

As any of our patients in these communities will tell you, being diagnosed with a rare disease can be a frustrating experience. Aside from dealing with the condition itself, there’s the lack of public awareness, a lack of research investigating your condition, and a sense that you are being “lumped in” with a similar disease because your community doesn’t have the critical mass to merit its own attention.  These new percentile curves for PLS and PMA patients demonstrate the value and power of openness.  By sharing their health data in an open fashion, patients are providing new insights that are changing how we think and act when it comes to these very rare conditions.

Note:  A potential limitation of these curves is that they represent the outcomes for patients that are members of PatientsLikeMe and may not be generalizable to the entire population; we are working hard to better understand and correct for the biases in our population and data. As the size and longevity of each community increases, we will be in a better position to address these issues.

PatientsLikeMe member pwicks PatientsLikeMe member tvaughn