Come for the donuts, stay for the data!

(data)modeling--how much crap can you take?

10/15/2018

A lot of modeling is how much crap you can take--Lauren Hutton

My first gig working in pharma was a lesson in acronyms. I had a 12 month contract at GSK and I learned an exponential amount. Everyone used verbal shorthand and I jotted the scattered letters down to look up later.

I think it softened my skin. I have no problem asking the questions. The biggest surprise is how few of us challenge the answers.

Although I have experience as a bench scientist, academic writer at a large academic medical center, as well as a lot of agency work--the added layer of pharmaceutical experience was the real hat trick.

Starting my data journey, I often heard statisticians discussing data models as if they were a thing I could pull from a shelf. They sound vetted and official don't they?

A data model is quite simply an "accumulated set of discovered relationships" within a dataset. Make no mistake the models are tested, but if sufficient "target variables" or "class labels" are not identified and considered, you can imagine the limit of insights to be gleaned from analyses.

Approached without care, data mining can reproduce existing patterns of discrimination, inherit the prejudice of prior decision makers, or simply reflect the widespread biases that persist in society.

It can even have the perverse result of exacerbating existing inequalities by suggesting that historically disadvantaged groups actually deserve less favorable treatment.--Big Data's Disparate Impact Barocas & Selbst

Working in health policy and accessing databases filtered by race, economics, and a wide variety of demographic target variables--it is critical to consider all of the "possible values of the target variable into mutually exclusive categories".

A 360 perspective is imperative while data mining. The ability to target variables specified and parsed to avoid systematically disadvantaging protected classes leads to actionable insights. It is a necessarily subjective process but can be improved with data literacy and applied to non-binary classifications.

Here is an important graphic from The Senate: Affirmative Action for White People published in the NY Times--written by David Leonhardt. Civics lesson rewind: residents of Washington, D.C., and Puerto Rico do not have voting representation in the Senate.

Right now, about four million American citizens have almost no congressional voting power, not even the diluted power of Californians or Texans. Of these four million people — these citizens denied representative democracy — more than 90 percent are black or Hispanic.

They are, of course, the residents of Washington, D.C., and Puerto Rico. Almost half of Washington’s residents are black, and nearly all of Puerto Rico’s are Hispanic.

“Big Data is like teenage sex: everyone talks about it, nobody really knows how to do it, everyone thinks everyone else is doing it, so everyone claims they are doing it.” – Dan Ariely

Walking down New Hampshire Avenue on my way back to my hotel, I stumbled upon the Democracy Tree. It is nestled between the roots of this majestic tree. Here is what it says:

This Tree is dedicated to the more than half million veterans, taxpayers, and citizens of the District of Columbia who, despite fighting in foreign wars, paying their full measure of taxes and continue to have no voting representation in the Congress of the United States of America.

"Taxation without Representation is Tyranny."
Foundry Democracy Project, Foundry United Methodist Church

The results are pretty outrageous. The Senate gives the average black American only 75 percent as much representation as the average white American. The average Asian-American has 72 percent as much representation as a white person. And the average Hispanic American? Only 55 percent as much. That’s right — the structure of the United States Senate treats a Hispanic citizen as only about half as important as a white citizen.

My point in all of this is surprisingly not political. The illustration here is to show that technology and advancements in our Big Data culture are forgetting the nuance of machine learning, AI, and algorithms--the insights are only as good as the data we feed into them.

"Data is frequently imperfect in ways that allow these algorithms to inherit the prejudices of prior decision makers."

The full quote below is an important example of how stewards of data (and aren't we all) have a unique responsibility to be thorough and informed as we explore and mine data. Garbage in, garbage out but worse yet--discrimination in, discrimination out...

Advocates of algorithmic techniques like data mining argue that these techniques eliminate human biases from the decision-making process. But an algorithm is only as good as the data it works with.
Data is frequently imperfect in ways that allow these algorithms to inherit the prejudices of prior decision makers. In other cases, data may simply reflect the widespread biases that persist in society at large. In still others, data mining can discover surprisingly useful regularities that are really just preexisting patterns of exclusion and inequality. Unthinking reliance on data mining can deny historically disadvantaged and vulnerable groups full participation in society.
Worse still, because the resulting discrimination is almost always an unintentional emergent property of the algorithm’s use rather than a conscious choice by its programmers, it can be unusually hard to identify the source of the problem or to explain it to a court.--Big Data's Disparate Impact Barocas & Selbst

Being the one that knocks...

10/12/2018

I know, I know, you didn't ask but I have some advice for you. Before you attend the conference or join the organization--think about who is being served by your attendance or membership. Are you "picking yourself?" as Seth Godin would say or are you part of someone else's narrative? I have joined organizations like the American Medical Writers Association but got very little in return. You become the product as they sell your data for a profit--I know--who doesn't anymore.

My pivot toward value and independence came when I found the right organizations. Most of them are focused in journalism so the art of storytelling and data literacy began to emerge.

I learned this lesson a single pain point at a time. When I didn't pick me, I ended up writing client biased content orchestrated by agency x, pharma company y, or medical education company z.

Earning press credentials provides access to medical and industry conferences--for free. When I was forced to tell sponsors what I was going to write about--I informed them I had no idea. Until I enter the space and have conversations, listen to the lectures or keynotes, or participate in the Q&A how can I possibly know what to write about? But that is how most journalism platforms work. You pitch or get pitched and then get to work. I reject this bias. I pick me. Or better yet, I pick you.

I reflected on this recently at a "meet and greet" for new members of the National Press Club in DC. After recent years of finally picking myself, I was invited to become a member. I think of the familiar creed among journalists,

Afflict the comfortable and comfort the afflicted

I now write about healthcare and all the different moving parts. Think of the piles of information and sort them by health policy, health economics, health law, clinical research, or whatever web that wipes across your mind.

We are taught to follow algorithms. Write the manuscript, book, article, or report as instructed. For a while, we comply, until we don't. We finally realize that you can't win by being more average than your competition. Better at following the blueprint. Think about your education or perhaps you have children navigating the educational system.

Watch them dutifully follow rubrics, guidelines, outlines, and doing what they are told. Our educational system is obsolete. We don't need these skills in modern society. We need creators, inventors, innovators--we need insights. If you have an idea of something you want to see in this world--listen to the short podcast by Seth Godin. Tag. You ARE it...

Well-meaning colleagues at the Press Club were excited to introduce me to members, experts in their fields, communicators, all to network--because I mentioned that I work for me. But it was weird. They were guiding me toward mentors. I reminded them that not all who wander are lost. I don't want a job. I am the job.

Who are you talking to right now? Who is it you think you see? Do you know how much I make a year? I mean, even if I told you, you wouldn't believe it. Do you know what would happen if I suddenly decided to stop going into work?

A business big enough that it could be listed on the NASDAQ goes belly up. Disappears! It ceases to exist without me. No, you clearly don't know who you're talking to, so let me clue you in. I am not in danger. I am the danger. A guy opens his door and gets shot and you think that of me? No. I am the one who knocks!--Walter White, Breaking Bad

Where I am headed next...and hope to see you!

Or reach out @datamongerbonny for scalable talks on data literacy, data sourcing, governance, or building your own data pipeline...

Knowledge is power but information is liberating

10/6/2018

The title of today's post is attributed to a statement made by Kofi Annan about education. On the surface, what could be more anodyne than knowledge right? But what if a large percentage of the collective, narrowly defines information based on a single perspective?

The remaining part of the quote, "Education is the premise of progress, in every society, in every family" might be the key element. I want to encourage you to pursue education daily. Many peers, mid-career or beyond, tuck into comfortable cozy careers either in a subject niche, algorithmic informed writing habit, or simply decide to focus on regulatory and compliance template driven writing. I am here to tell you--those days are waning. Big data is here and has invited CRISP-DM along for the ride. Briefly, the 6-phases of cross-industry standard process for data mining include understanding your data, preparation, modeling, evaluation, deployment, and business understanding.

You might be thinking--hey wait a minute--I am a writer not a data person. Okay. But here comes the hard no. If you aren't data informed or aware, what are you actually writing? Opinions?

The Government are very keen on amassing statistics - they collect them, add them, raise them to the nth power, take the cube root and prepare wonderful diagrams.--Josiah Stamp

Last week was a return to travel, workshops, and a smattering of data literacy work from the setting of a local conference on interdisciplinary statistics and combinatorics. I had grounded myself for a few months. Opting to step out of the frenzied and chaotic world of conference attendance and ask myself an important question--who is being served by me being in the audience?

Am I able to bring information immediately applicable at the point of care or am I listening to "experts" proclaim mind-numbing self-promotion and shameless marketing? Media credentials can get you front row access to most conferences but I would also argue--beware the chattering masses--they have heterogeneous motivations and goals. It is quite all right to return to square one--learn what you didn't know, you didn't know...

Thursday afternoon I traveled to North Carolina State University to attend a 2-hour workshop R for Document Creation (R Markdown). Because I am handling large datasets and would like the data to travel with the visualizations I am learning how to make this happen. I am learning R (with Python refreshers to follow) to expedite working with non-proprietary healthcare data. The datasets are large and often in need of cleaning and preparation for analysis. Check out your local universities. Free workshops are a great way to add a few new "shiny" tools to your skillset.

Here is an example of how Edward Tufte uses R markdown to prepare margin figures in his books and writings...The beauty of R--or at least for me--is the simplicity of the scripts. You can name them and save them for quick updates or application to additional data sets.

Using or learning R has been impactful and instrumental in allowing me to approach data visualization with a question informed by multiple datasets but not limited by interoperability or access. When you "knit" the code--voila the image will be published with the code either displayed or neatly hidden away--your choice.

This weekend, I wandered over to a local conference being held at an alma mater, UNC-Greensboro. Invited as media I scanned the conference brochure looking for sessions of interest (and was not disappointed).

I will be writing more in depth about the sessions but for now the familiarity of Simpson's paradox and refinement of my ability to interpret gene signature data were highlights.

You can amass data. That is the easy part. Big data lakes host a wide variety of potential information but you will need buckets--S3 or Hadoop clusters to organize. You will need pipelines to your data in order to begin the process. Upstream insights to confirm the right data is being generated (garbage in, garbage out), all 6 steps of CRISP-DM, and a governance strategy to protect your data and control access.

Follow along for information and guidance--or reach out for a hand. You never know what you might find swimming beneath the surface...

Clarity in the eye of the beholder...

9/20/2018

A recent article, STATISTICAL PARADISES AND PARADOXES IN BIG DATA (I): LAW OF LARGE POPULATIONS, BIG DATA PARADOX, AND THE 2016 US PRESIDENTIAL ELECTION reminds me of a great definition of statistics, "Principled thinking and methodology development for dealing with uncertainty." My intent is certainly not to be partisan but to highlight data that could just as easily been rendered in a behemoth healthcare data set. IPUMS describes the process well, "Think of census data needing to be cleaned, merged with new data, editing routines developed, and millions of strings coded into useful classifications. The data are far too large for manual inspection, requiring efficient data analysis and scalable approaches including machine learning."

The article is technical but can be parsed into tangible little morsels. Here are a few that interest me. Large-sample asymptotic is basically the approximation or limit of a graph--you have probably heard of the Central Limit Theorem and it has a role on the limits of Big Data. So I don't go down the rabbit hole of technical jargon this is a great analogy from the article (you should read it). Think of statistics and dealing with uncertainty--the author uses food metaphors--they help.

Fast food will always exist because of the demand—how many of us have repeatedly had those quick bites that our doctors have repeatedly told us stay away from?

But this is the very reason that we need more people to work on understanding and warning about the ingredients that make fast food (methods) harmful; to study how to reduce the harm without unduly affecting their appeal; and to supply healthier and tastier meals (more principled and efficient methods) that are affordable (applicable) by the general public (users).

I will bullet the most important findings in the article taken mostly verbatim:

difference between the sample average and the population average is the product of three terms: (1) a data quality measure, (2) a data quantity measure, and (3) a problem difficulty measure
the most critical—yet most challenging to assess—among the three is data quality.
When combining data sources for population inferences, those relatively tiny but higher quality ones should be given far more weights than suggested by their sizes.
data quality must be a relative one, and more precisely it should be termed as data quality for a particular study. This is because any meaningful quantification of data quality must depend on (1) the purposes of the analysis—a dataset can be of very high quality for one purpose but useless for another; (2) the method of analysis (e.g., the choice of sample average instead of sample median); and (3) the actual data the analyst includes in the analysis
Big Data Paradox-- The bigger the data, the surer we fool ourselves.

Fig 4. (please go to article for details) comparison of actual vote shares with 2016 Cooperative Congressional Election Study (CCES) estimates across 50 states and DC. "Color indicates a state’s partisan leanings in 2016 election: solidly Democratic (blue), solidly Republican (red), or swing state(green). The left plot uses sample averages of the raw data (n = 64,600) as estimates; the middle plot uses estimates weighted to likely voters according to turnout intent (estimated turnout nˆ = 48,106);and the right plot uses sample averages among the subsample of validated voters (subsample size, n = 35,829)."

Observe the only voting regions where the confidence intervals (barely) cover the actual results (Trump Data). And it should provide a clear warning of the Big Data Paradox: it is the larger turnouts that lead to more estimation errors because of systemic (hidden) bias, contrary to our common wisdom of worrying about increased random variations by smaller turnouts.

This article is one you may need to read and re-read several times. I find it important as we enter the age of personalized medicine. You can see here--even if the mathematics underlying the problem are complex--individualized predictions are approximations.

Because each of us is unique, any attempt to “personalize” must be approx- imative in nature. But this is a very different kind of notion of approximation, compared to the traditional large-sample asymptotics, where the setup is to use a sample of individuals to learn about the population they belong to.

In contrast, individualized prediction is about finding a sequence of proxy populations with in- creased resolutions to learn about an individual. This leads to an ultimate challenge for Statistics (and statisticians): how to build a meaningful theoretical foundation for inference and prediction without any direct data?--Xiao-Li Meng, Whipple V.N. Jones Professor of Statistics, Harvard Faculty of Arts & Sciences Editor in Chief of the Harvard Data Science Review

Why do so many of us view statistics and data literacy as something we just aren't built to comprehend?Sal Khan from Khan academy hits the nail on the head. If you are like me and annoyed by everything of value being tucked behind a paywall--his motto is, "You can learn anything. For Free. For Everyone. Forever."

Finally I now know why I am able to grasp advanced statistical principles (at least the big ideas) readily in my modern day work flow. He summarizes below in his brief and worthwhile talk below--only about 10 minutes but BOOM it is worth it. In a nutshell, when we are self-directed and learning or re-learning a skill or subject we tend to focus on what we don't know before moving on.

In a traditional academic model, we group students together, usually by age, and around middle school, by age and perceived ability, and we shepherd them all together at the same pace. And what typically happens, let's say we're in a middle school pre-algebra class, and the current unit is on exponents, the teacher will give a lecture on exponents, then we'll go home, do some homework.

The next morning, we'll review the homework, then another lecture, homework, lecture, homework. That will continue for about two or three weeks, and then we get a test. On that test, maybe I get a 75 percent, maybe you get a 90 percent, maybe you get a 95 percent. And even though the test identified gaps in our knowledge, I didn't know 25 percent of the material.Even the A student, what was the five percent they didn't know?

I never thought about the challenge of recalling academic courses and specializations in advanced Calculus, statistics, or programming languages. It was always about the test and let's face it, if you got an 85% on a test you were chuffed. Or at least I was. But think about the 15% of the material you didn't know? What if your car brakes were repaired only 85% of the way--would you consider it safe?

I thought I would share a few simple graphics I created to demonstrate clarity and what granularity can yield if we seek to understand what the small differences in a bar chart might mean if we are looking at the country as a whole.

The last graphic is of the electoral college votes--the only vote that matters in the US but you can see how additional clarity can once again be more insightful than reliance on a simple statistic about winning an election...thanks to Alberto Cairo for the image below.

Reach out for any help with data literacy or creating your own data stories--either in your professional or personal life--brainstorming is always free! twitter.com/datamongerbonny

data struts and frets his hour upon the stage...

9/2/2018

Out, out, brief candle! Life's but a walking shadow, a poor player That struts and frets his hour upon the stage And then is heard no more; it is a tale Told by an idiot, full of sound and fury, Signifying nothing.--Macbeth by Shakespeare

As similar drugs hit the market, the pundits race to see where the profits are likely to be highest. Case in point look no further then immuno-oncology drugs or other categories within personalized medicine.

"For now, sales are a useful metric fo starting and staying on a checkpoint (Inhibitor)"

I would argue that patient outcomes like "overall survival" will be the metric that delivers. We struggle to reconcile patient centered care with the aggregated data that leads to FDA approval. But meanwhile, premature media hype of early phase prognostic factors and clever pre-launch hype have become the harbors of safety investors have in the crosshairs.

Commercialization of disease has become a huge profit motive for shareholders. I have no problem if you are owning your decision to hedge the bets in favor of expanding market shares but I also see where patient as a buzzword has become like an awkward "comma" in the overall story. "What is that doing there and is it really necessary--It disturbs the flow of the narrative".

My shareholders expect me to make the most profit. That's the ugly, dirty truth.--Martin Shkreli

I have been painfully unsuccessful in talking about pharmaceuticals in a vacuum that does not include health policy, health law, health economics, and clinical research data. It bugs some people. I work as a consultant that views my role as being truthy and delivering perspectives often over-looked in the Google economy. The majority of calls I am on include a C-suite type masquerading as an expert without ever exhibiting doubt in the science or wanting engagement on other perspectives.

I recognize discussion topics from the first pass on what loads on page 1 of an online search. What I don't hear is the granular discussions from well-designed trials that add a depth to pre-clinical decisions or moiety designs.

Attempts to vilify the pharmaceutical industry without transparency of health law, policy, and economics underlying the framework continue to be short sighted. In our big data lives, there are now infinite data libraries and resources for query and analyses. Was it Martin Luther King that warned us all, "Nothing in the world is more dangerous than a sincere ignorance and conscientious stupidity"?

Upstream from diseases we can discover how laws influence health. No matter what side of the aisle you are on--changes to federal law, reversal of executive orders, changes to federal agency policy (ask yourself how Alex Azar, Health and Human Services Secretary, ended up with jurisdiction over immigration) and appointment of judges impact public health.

The Southeastern Association of Law Schools 2018 conference featured a session on health law:

WORKSHOP ON HEALTH LAW
Healthcare in the Era of the Trump Administration
The 2016 election has already ushered forth dramatic changes to our healthcare system, in efforts to repeal and replace the Affordable Care Act, in deregulation, in executive orders that alter the effects of the ACA, and in numerous other ways. This panel will focus on the state of healthcare in America in the era of the Trump presidency. What are the enduring effects of the Trump presidency on the state of healthcare delivery and healthcare finance (both in private and public insurance programs) now and into the future?
Moderator: Professor Deborah Farringer, Belmont University College of Law
Speakers: Professor Jennifer Bard, University of Cincinnati College of Law; Professor Zack Buck, The University of Tennessee College of Law; Professor Nicole Huberfeld, Boston University School of Law; Professor Nicolas P. Terry, Indiana University, Robert H. McKinney School of Law

I recommend listening to This Week in Health Law podcast if you want to access a wide variety of legal scholar discussions outside of the typical journalistic soundbites and tantalizing ledes. These are a few of the graphics from the event. The broad discussions include the role of trade tariffs on the cost of medical equipment, the fight against global guidelines on livestock antibiotics, suggested bans on CDC employees using words like "evidence-based, science-based, vulnerable, entitlement, diversity, transgender, and fetus", the phrase "climate change", sexual discrimination and Title VII protections, use "religious liberty" to deny medical care, and on and on...

Harvard scientists say current and proposed environmental policies could lead to an extra 80,000 deaths per decade. Think of losing the entire audience from the MetLife Dome --where the New York Giants and Jets play--every decade because of changes in policies that impact our clean air and water. Requiring raw data from landmark cases to be released to the public or risk overturning of established science.

Letter from Harvard leaders to EPA's Andrew Wheeler on proposed science policy by Emily Atkin on Scribd

Context is king. We will not be able to out "Innovate" our way out of global and national priorities that minimize the role of social determinants, health policy, social policy, economics, and the environment.

time to get our hands dirty...

I for one do not accept Shakespeare's existential nihilism from Macbeth. We just need to look at more data. Dirty hands get things done.

Find more data--not just the data that supports our objectives, bottom line, or shareholder interests.

Minding my own business

8/27/2018

A fellow trail runner suggested the title for a new post. When you are out on the trails you tend to share things and admittedly you likely think things are funny that may or may not be. Regardless, I cleaned up the actual suggested title and decided to go with the sentiment. How are you running your business? It can be your own small digital media company or the business of you within the confines of an employee/employer establishment.

There is never a lack of listicles published to help you decide to stay an employee or branch out on your own into the wild and woolly world of entrepreneurship. I was sort of pushed. Small companies are at risk of insolvency from external pressures and internal--sometimes both.

You can be brave out on your own but sometimes its easier with an established cast of characters. We know the gossip, the BS artist, the person that will always deliver during crunch time, the data person--what do you do if they are all you? Finger pointing and hushed laughter lose the luster when you have no one but yourself to blame. I was recently asked to write something described as "traditional". What does that even mean? I think it was supposed to be a pitch--Our Product is Great--Everyone Else Stinks--Our Product is the Best. I have many tools I rely on and couldn't imagine working without but lets face it, there is always room for improvement.

What drives your curiosity when you read about R&D driving up healthcare costs? How is R&D being defined? How much research are we talking about or is it primarily mergers & acquisitions and patent protection? How much of our tax dollars have funded the "Innovation" being heralded as the cost of doing business in our new global economy?

I detest organizations in the news business that park everything of interest behind a paywall. Even more when the discussions are around important considerations in healthcare. News we all need to know to make informed critical decisions. I know, I know, its all about making a buck but what about when it isn't?

It reminds me of the news headline on the evening news (remember the evening news?)

You Won't Believe What Household Item Will Kill You in About 10 Minutes.--News at 11.

I remember when the news was for all of us. Not just those of us paying for premium access. Don't worry. It makes me want to work harder. To ask more questions and question more answers.

I go where my curiosity leads me. Beth Macy wrote a fabulous book about the opioid crisis, Dopesick. Dealers, Doctors, and The Drug Company That Addicted America.

Not because it was in the news media but because she noticed something not being discussed when she was writing about the furniture industry--what was happening to communities in the wake of offshoring.

I always feel a little preachy when finishing a workshop presentation. The audience always seems surprised regarding the cost of limited data literacy. Once upon a time it became okay to roll our eyes around mathy topics. That time is gone.

When we can't go right to the source and decipher the statistical modeling or data sources we are at the whim and fancy of people that own the knowledge. I don't remember who said it but it went something like this: Discovery isn't followed by someone yelling EUREKA! It is usually the quieter, hmmmm...Isn't that interesting?

Those with the knowledge control how much of the truth valve to open.

What if it is only a small dribble?

"Liberty, prosperity" and streptomycin?

8/11/2018

I am a New Jersey gal. Even when I think I am a California gal or North Carolina gal--there are some things embedded in identity that you can't shake. For example, Bruce Springsteen sang quite convincingly from the stage that I should "get out while you're young"--and I did.

On a long recent drive to NYC my husband (Steve) mentioned New Jersey is seeking to adopt a state microbe. I had no idea. And then he shares NJ is only the second state to put it up for a vote. The first state? Oregon and the winning microbe--was brewers yeast. Saccharomyces cerevisiae proudly represents the many small craft breweries across the state.

Our likely winner here in the Garden State will be Streptomyces griseus, the producer of streptomycin.

Streptomycin was discovered by Albert Schatz, Elizabeth Bugie and Selman Waksman in January of 1944. It was truly remarkable in the breadth of pathogens that it attacked, including Vibrio cholerae (cholera) and Mycobacterium tuberculosis (TB), against both of which penicillin, the only other antibiotic at that time, had no effect. The discovery of streptomycin radically changed public health.

This producer of streptomycin (Streptomyces griseus), which was discovered in New Jersey soil and also isolated from the gullet of a healthy New Jersey chicken, was developed at the Department of Biochemistry and Microbiology, Rutgers, the State University of New Jersey, where its remarkable antibiotic properties were recognized. It was used in the development of large scale fermentation, a novel concept at that time. There had been war-time pressures to increase the production of penicillin via deep tank submerged fermentations. This was successful, evidenced by the general availability of penicillin on D-Day, June 6, 1944. The pressure for enhanced penicillin production continued. George Merck, in assessing needs for penicillin and those for the newly found streptomycin, overrode Merck's scientific board, which was focused only on penicillin, and committed extra funding for a major new project: the production of streptomycin.

Thus streptomycin joined penicillin in the development of the pharmaceutical industry in New Jersey. Within four years of the discovery of streptomycin in 1944, the initial yields in the milligram range were ramped up to 3,000 kilograms per month, a million-fold increase. This enabled animal and human trials by Drs. William Feldman and Corwin Hinshaw, at the Mayo Clinic in Rochester, MN.--Rutgers School of Environmental and Biologic Studies

I love a quote by NPR host, Ailsa Chang "That microbe has saved lots of lives, including Senator Bob Dole and Beatles drummer Ringo Starr. Now, it has one more challenge - becoming official state microbe."

I have been thinking about interesting data and facts that send us off in pursuit of more context and knowledge. Motivated by the abundance of non-proprietary data free for us to access in our personal and professional lives--I began teaching and writing about data literacy. Watch this space for a new creative project to help all of us use data more effectively and give voice to stories yet to be heard...

Nimble solutions for big data...

8/2/2018

What a busy week. I am taking a refresher R course, Introduction to R for Journalists: How to Find Great Stories in Data. I also hop in and out of the Data Science Specialization created by The Johns Hopkins over at Coursera. Admittedly the R Programming course is quite intensive but if you make it out the other side you will be in masterful shape to tackle large datasets in relatively no time at all. I highly recommend the 10 course specialization but admit it is hard to stay focused while juggling client work and travel.

But here is why it is important. If you are safely ensconced in Excel spreadsheets and tables--good luck. That data endeavor is manual and labor intensive. You need to rebuild the ship every time you decide to set sail. Not the best use of your time. Data projects rely on open source data and also a seemingly infinite number of non-proprietary data floating around the web. Data brokers have created streamlined solutions by cleaning the data for you and in many cases combining datasets to provide longitudinal analyses.

A little elbow grease though, and you are able to write a little code to update or tidy your date on the fly. As someone who did this manually to create data visualizations for clients--the juice is worth the squeeze. I code in R and Python and the living is easier...

This is also the first week in several, where work won out over running on the trails and training for an ultra-event. Sometimes only able to secure a few hours in the pre-dawn, I use the time to re-enter work life with the scale tipped back toward work. I don't know how you stay on your professional toes but I would be sunk without podcasts whispering insights and ideas into my ears.

Perhaps not a typical vacation read, but The Open Revolution: Rewriting the Rules of the Information Age introduced me to the work and writings of Rufus Pollock an economist and founder of Open Knowledge International.

In this new world, intellectual property is intellectual monopoly. Monopolies are unjustified and unjust, dangerous both to our economies and our societies. We need new rules for this new, digital world: rules appropriate to the information economy; rules that provide ways to reward innovators and creators whilst preserving fairness and freedom, and which give everyone a stake in our digital future.--The Open Revolution: New rules for a New World by Rufus Pollock

I like the powerful analogy of baking a cake. If the ingredients are locally available or even in your pantry--well done you. But what if you had to drive to the farm and wait for the flour to be milled, eggs to be produced, etc. Well a data or insight "cake" has similar challenges. Sources of friction--legal issues, data quality, and data logistics are all financial as well as temporal barriers to efficiency.

The Frictionless Data Field Guide is for those of us working with data of all stripes. I found it an erstwhile companion to my work in R as a guide to workflows, data collections, data sourcing, validation, getting data out in the world, and improving data publishing. For the sake of transparency I made a mess of my desktop and dropbox files as I manipulated huge amounts of data into repositories and visualizations. After each project I dutifully collected the raw data and packaged the workbooks for clients but needed to devote not a trivial amount of time creating order from the flotsam. The workflow has made my data sources ready for the next query and accessibly organized.

In the video below, Rufus explains how the containerization of shipping automated reduced costs and increased efficiency by an order of > 1000%. Think of data packages as containers for data--once we have standardized data "containers" we have tools to validate, store, search, import, and export data.

The Economist: Babbage podcast is a nice visit with science and technology and an easily assimilated brief audio snap. The namesake of Charles Babbage the "19th-century polymath and grandfather of computing" presents a potential seed of a solution for drug pricing.

In continuing the Rufus Pollock theme--he appears momentarily but discusses the tensions around drug price transparency and monopolies.

A firm supporter of innovation, he recognizes that the big player in tech isn't necessarily the associated gadgetry but the basic rules around ownership. Enterprise, free markets, innovations and inequality depend on rules we combine.

We are operating as if monopolies and patent rights are a single tool and the only solution to rewarding creators for their investments in R&D. I admire the "out of the box" thinking around better solutions to improve access to life-saving drugs.

The book and podcast distinguish patent rights from property rights.

The problem arises when patients will die without access to drugs. If they are priced prohibitively to cover R&D and the marginal cost of manufacturing with one price--fewer drugs are sold and patient outcomes may be quite draconian. But what if, as Rufus suggests, there is a two part payment model--cost of goods has the fixed cost of R&D and cheap marginal cost for medicines relative to R&D.

Innovators can invest or pay for two things separately. Subscription fees pay for fixed cost and manufacturing costs are calculated separately. An enumeration model provides access through a fixed fee to cover R&D and money is allocated to a fund to pay for which drugs get used and how effective they are. No need for government to distribute--only to make sure money is collected. An actual plausible model not trying to be crammed into an outdated framework like US healthcare ecosystem.

I have to admit, it took some time to snap out of my holiday routine of long runs in Montauk and leisurely attention to projects needing a little "kicking down the road". But as we all know--there is much work to be done. I am a new member of the National Press Club and ideas for stories are buzzing around my head constantly. The newly refreshed skills in R programming language help to swat away the time factor. Yes, you need to focus and do the work up front--but once you create your datasets and data frames--you are frictionless...

“The best thing to do with your data will be thought of by someone else”-Rufus Pollock

And soon, too soon, we part with pain, to sail o'er silent seas again--Thomas Moore

7/28/2018

I have a fondness for Irish poets. When I was reading the beautiful poem, PAIN, by David Whyte I was reminded of Thomas Moore and his benediction like quote about the ubiquity of pain--an unavoidable side effect of being alive.

Weirdly it applies to my work as a data scientist. To truly quantify and qualify pain I seek to calibrate definitions. Industry often treats pain like a monolithic homogenous entity and it is certainly NOT. I pick variables and attributes hoping to measure pain but know that without the context and cipher to translate--It is meaningless.

I hope you will take the few moments to read the poem. If you work with patients in any capacity or are tasked with research methodology in questionnaire design or work with healthcare providers--you will appreciate how a simple pain scale serves no one.

I was fortunate to hear David read his poems live at the first ever On Being retreat held beneath the beautiful redwoods in California. He has a hauntingly beautiful tone and cadence. Listen to him from the TED stage. David Whyte has many beautiful books of poetry but one of my favorites is Consolations: The Solace, Nourishment and Underlying Meaning go Everyday Words. The poem Pain is one of the 52 words selected as a "doorway into the underlying currents of human life"...

I for one don't think we live enough in the space of what a word actually means. I recall robust debates between stakeholders when collaborative efforts were stalled as we intently tried to define such familiar terms such as "middle-class" or "value" in a way that allowed merging of large datasets.

Pain is a big one. When we examine addiction culture, rising suicide rates, and patient outcomes across a wide variety of chronic conditions are we certain we are measuring the same thing?

Experiencing real pain ourselves, our moral superiority comes to an end; we stop urging others to get with the program, to get their act together or to sharpen up, and start to look for the particular form of debilitation, visible or invisible that every person struggles to overcome. In pain, we suddenly find our understanding and compassion engaged as to why others may find it hard to fully participate.--David Whyte

Please share, comment over on twitter, or donate...thank you!

Mendelian randomization and casual inference oh my...

7/2/2018

I rely on podcasts to keep me entertained on long trail runs. Initially it was by necessity as there are only so many hours in a day. If you are going to lace up the sneakers and be off grid during a busy work week--the juice better be worth the squeeze. What started as a diversion for entertainment quickly escalated into an academic exploration into a wide variety of topics.

Candidly, rigorous and reliable sources of information are tainted with profit motives, pay-walls that limit information exchange, bias, and large egos. For these reasons and many others I have relied on the written insights of Peter Attia, MD. I have written about him before, How Should We Discuss Emerging Drugs for Patients with High Cholesterol and read everything he writes.

Peter has a new podcast. It is long, detailed, and perfectly technical for my taste but you may either be overwhelmed or not the sort that can appreciate a long listen--but you should. It is brilliant. The show notes alone will make you better informed. I will share a few highlights here but if you have commute time or schedule meaningful education into your work life--go all in.

I worked with a brilliant cardiologist for several years and was introduced to critical thinking, potential bias in peer review, and from time to time--the limits of statistical analysis in the absence of qualified interpretation. In fact, the name of this blog was pulled from my days working with research fellows required to present their data in a journal club format. I often stated, Come for the Donuts, Stay for the Data as we shared Krispy Kreme donuts outside the lecture hall.

Consider this a warm morsel right from the oven--dig in...

If you are going to dive into only a few of the topics discussed I think your first hop should be over to The early years of lipoprotein research: from discovery to clinical application.

Many recent studies, however, have under-scored the importance of considering the heterogeneity among LDL and HDL particles, e.g., their quality and number, and their variable association with CVD risk.

Research evolving across a wide variety of therapeutic areas continues to point toward metabolic derangement as a major player in chronic disease.

Discussions of Mendelian randomization spurring engagement of genetic variants as tools for examining modifiable exposure in observational studies are intriguing. Particularly relevant when looking to illuminate the role of limiting cholesterol synthesis vs. focusing on receptor integrity.

Think of PCSK9 inhibition and what we know about the blunted exposure of clinical trial efficacy compared to lifetime exposure in individuals with a PCSK9 genetic mutation showing changes in hyper- function or hypo-function (lowering LDL).

You may recall from your high school science that Mendel was a 19th century monk that discovered pathways of genetic inheritance by studying peas. An assumption of randomization provides the ability to study a single genetic variant or collection of variants associated with a biomarker to test causality of a disease process. Differences observed with genetic variants demonstrate that life long exposure predicts CVD risk more robustly than clinical trials because the magnitude of effect during investigational trials is limited in exposure.

I am working my way through the resources listed for this recent podcast. I have always been puzzled about the demarcation of disease by organ system and specialties. It may be the reason that we are unable to decipher meaningful advancements in Alzheimer's disease. Reduced plasma desmosterol-to-cholesterol ratio and longitudinal cognitive decline in Alzheimer's disease provides an investigation into an association between plasma desmosterol-to cholesterol (DES/CHO) ratio in patients with Alzheimer's disease and longitudinal cognitive decline. You can see in the graphic from the study; the control (blue), mild cognitive impairment (green), and Alzheimer's disease (AD) (red) findings.

Plasma DES/CHO is reduced in AD patients and correlated with Mini-Mental State Examination (MMSE) score.

"The longitudinal analysis revealed that plasma DES/CHO in AD patients shows a significant decrease at follow-up intervals. The decline in plasma DES/CHO is larger in the AD group with rapid progression than in that with slow progression. The changes in plasma DES/CHO significantly correlated with changes in the MMSE score."

These findings seem to support DES/CHO as a potential longitudinal surrogate marker associated with cognitive decline outside of protein deposition in the brain. I will post more over at Alzheimer's disease: The Brand once I read the resources posted by Peter Attia.

I hope you will head over to the podcast for this important episode. Here is the bulleted list from the podcast website.

The pathogenesis of atherosclerosis;
How early atherosclerosis begins;
Ron’s motivation for getting into lipidology;
How reading an article series in the NEJM in 1967 had a profound impact on him and his career;
The “battle” between LDL particle size and particle number;
The use of statins;
The role of chronic inflammation in atherosclerosis;
Why niacin may have been unjustly dismissed as a therapeutic option;
The HDL paradox: why drugs that raise HDL-C seem to elevate (or have little impact on) heart disease risk;
Mendelian randomization: nature’s randomized trial;
How PCSK9 inhibitors work and why they may be underutilized;
Lp(a); and more

<<Previous

Forward>>

Browse the archive...

Follow @datamongerbonny

Thank you for making a donution!

In a world of "evidence-based" medicine I am a bigger fan of practice-based evidence.

Question the quality of the evidence. The motivation for disseminating the evidence.

Who stands to benefit the most from its uptake?

Remember the quote by Upton Sinclair...

“It is difficult to get a man to understand something, when his salary depends upon his not understanding it!”

Tweet to @datamongerbonny

Where I am headed next...and hope to see you!

time to get our hands dirty...

Data and Donuts has a free weekly newsletter!

Categories