Skip to content
  • About
  • Contact
  • Contribute
  • Book
  • Careers
  • Podcast
  • Recommended
  • Speaking
  • All
  • Physician
  • Practice
  • Policy
  • Finance
  • Conditions
  • .edu
  • Patient
  • Meds
  • Tech
  • Social
  • Video
    • All
    • Physician
    • Practice
    • Policy
    • Finance
    • Conditions
    • .edu
    • Patient
    • Meds
    • Tech
    • Social
    • Video
    • About
    • Contact
    • Contribute
    • Book
    • Careers
    • Podcast
    • Recommended
    • Speaking

Big data equals big challenges for prostate cancer

Leonard G. Gomella, MD
Conditions
September 24, 2016
Share
Tweet
Share

The term “big data” is a favorite mantra of health care today. Big data sets are starting to drive much of what is done in medicine including directing research, drug development, clinical pathways, insurance coverage and public opinion. The official definition of “big data” in health care is subject to interpretation by different sources. One dictionary defines big data as “data of a very large size, typically to the extent that its manipulation and management present significant logistical challenges.” According to another definition, health care data are considered “big” if it represents many more subjects than a typical randomized clinical trial (tens of thousands or more) and/or includes a broad range (hundreds or more) of clinically relevant patient and provider characteristics.

But big data sets while large and typically considered robust, are not infallible. Two recent high-profile prostate cancer data sets raise concerns on how erroneous data may harm progress in this disease.

The Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute contains over 9 million entries spanning from 1973 to 2013, the most recent data available. While some might argue this may not fit the purist’s modern definition of big health care data, any mere mortal would likely consider 9 million entries “big” data. While not all SEER cases are prostate cancer, a significant portion of this information relates to prostate cancer directly or as a comparator to other cancers. One concept everyone can agree on is that any dataset relies on the accuracy of the data entered to make sound conclusions.

In late 2014 auditors identified a potential problem with PSA data entered in several years of the SEER database. The error related to the simple misalignment of a decimal point in how the PSA data were entered. Initially, almost 18 percent of the PSA data were felt to have been miscoded. The audit is ongoing, and while the error rate may prove not to be as high as first thought, the damage is done as PSA data have been pulled from the 2014 SEER update. More importantly, confidence in this well regarded large and mature dataset has been eroded.

The mere suggestion of inaccuracies in the revered SEER database sent shock waves across the academic and research communities. Many well-known researchers expressed concerns that previously reported studies, particularly in the controversial area of routine screening for prostate cancer, may have been flawed if they relied on this erroneous SEER PSA data.

Speaking of the prostate cancer screening controversy, much of the recent debate can be traced to the findings of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. The U.S. Preventive Services Task Force (USPSTF) relied heavily on data from this large trial to issue their 2012 recommendation not to screen for prostate cancer. With approximately 76,000 men, PLCO is the largest U.S. prostate cancer screening trial. Men were randomly assigned to annual PSA screening or standard care. The results showed no difference in prostate cancer mortality whether screened or not. However, a recent NEJM re-analysis indicated that up to 90 percent of men in the control arm had at least 1 PSA test before or during the trial resulting in contamination of the control group. With the lack of a true control group, this “big” PLCO trial seems to harbor erroneous prostate cancer data as well.

Modern medicine is poised to continue collecting massive amounts of information on all aspects of health care. Mining big data for research and health care policy determinations have been with us for some time and will assume an even greater role in the future. Bigger usually means better, but as these two prostate cancer experiences demonstrate, that might not always be true.

Leonard G. Gomella is chairman, department of urology, Sidney Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA, and editor in chief,  Canadian Journal of Urology, where this article originally appeared.

Image credit: Shutterstock.com

Prev

The yoga of surgery

September 24, 2016 Kevin 1
…
Next

Fibromyalgia: Medicine's F-word

September 24, 2016 Kevin 6
…

Tagged as: Oncology/Hematology

Post navigation

< Previous Post
The yoga of surgery
Next Post >
Fibromyalgia: Medicine's F-word

ADVERTISEMENT

More by Leonard G. Gomella, MD

  • Before ordering a genetic test, get to know GINA

    Leonard G. Gomella, MD
  • Is robotic assisted prostatectomy worth the added expense?

    Leonard G. Gomella, MD

Related Posts

  • Despite progress in cancer care, cost and equity challenges still must be addressed

    David M. Aboulafia, MD
  • Hormone replacement therapy is still linked to cancer

    Martha Rosenberg
  • We have a shot at preventing cervical cancer

    Lisa N. Abaid, MD, MPH
  • Obstruction of medical justice: How health care fails patients with cancer

    Miriam A. Knoll, MD
  • Using the Avengers to explain how cancer treatments work

    Jennifer Lycette, MD
  • Is social media a friend or foe of science?

    Michael Joyce, MD

More in Conditions

  • Why PSA levels alone shouldn’t define your prostate cancer risk

    Martina Ambardjieva, MD, PhD
  • Reframing chronic pain and dignity: What a pain clinic teaches us about MAiD and chronic suffering

    Olumuyiwa Bamgbade, MD
  • America’s ER crisis: Why the system is collapsing from within

    Kristen Cline, BSN, RN
  • Why timing, not surgery, determines patient survival

    Michael Karch, MD
  • Why psychotherapy works and why psychotherapy fails

    Peggy A. Rothbaum, PhD
  • How oral health silently affects your heart, brain, and body

    Charles Reinertsen, DMD
  • Most Popular

  • Past Week

    • Forced voicemail and diagnosis codes are endangering patient access to medications

      Arthur Lazarus, MD, MBA | Meds
    • How President Biden’s cognitive health shapes political and legal trust

      Muhamad Aly Rifai, MD | Conditions
    • The One Big Beautiful Bill and the fragile heart of rural health care

      Holland Haynie, MD | Policy
    • Why timing, not surgery, determines patient survival

      Michael Karch, MD | Conditions
    • Why health care leaders fail at execution—and how to fix it

      Dave Cummings, RN | Policy
    • How digital tools are reshaping the doctor-patient relationship

      Vineet Vishwanath | Tech
  • Past 6 Months

    • Forced voicemail and diagnosis codes are endangering patient access to medications

      Arthur Lazarus, MD, MBA | Meds
    • Why are medical students turning away from primary care? [PODCAST]

      The Podcast by KevinMD | Podcast
    • How President Biden’s cognitive health shapes political and legal trust

      Muhamad Aly Rifai, MD | Conditions
    • The One Big Beautiful Bill and the fragile heart of rural health care

      Holland Haynie, MD | Policy
    • Why “do no harm” might be harming modern medicine

      Sabooh S. Mubbashar, MD | Physician
    • The hidden health risks in the One Big Beautiful Bill Act

      Trevor Lyford, MPH | Policy
  • Recent Posts

    • Why point-of-care ultrasound belongs in every emergency department triage [PODCAST]

      The Podcast by KevinMD | Podcast
    • Why PSA levels alone shouldn’t define your prostate cancer risk

      Martina Ambardjieva, MD, PhD | Conditions
    • How to handle chronically late patients in your medical practice

      Neil Baum, MD | Physician
    • Reframing chronic pain and dignity: What a pain clinic teaches us about MAiD and chronic suffering

      Olumuyiwa Bamgbade, MD | Conditions
    • How early meetings and after-hours events penalize physician-mothers

      Samira Jeimy, MD, PhD and Menaka Pai, MD | Physician
    • Why medicine must evolve to support modern physicians

      Ryan Nadelson, MD | Physician

Subscribe to KevinMD and never miss a story!

Get free updates delivered free to your inbox.


Find jobs at
Careers by KevinMD.com

Search thousands of physician, PA, NP, and CRNA jobs now.

Learn more

Leave a Comment

Founded in 2004 by Kevin Pho, MD, KevinMD.com is the web’s leading platform where physicians, advanced practitioners, nurses, medical students, and patients share their insight and tell their stories.

Social

  • Like on Facebook
  • Follow on Twitter
  • Connect on Linkedin
  • Subscribe on Youtube
  • Instagram

ADVERTISEMENT

  • Most Popular

  • Past Week

    • Forced voicemail and diagnosis codes are endangering patient access to medications

      Arthur Lazarus, MD, MBA | Meds
    • How President Biden’s cognitive health shapes political and legal trust

      Muhamad Aly Rifai, MD | Conditions
    • The One Big Beautiful Bill and the fragile heart of rural health care

      Holland Haynie, MD | Policy
    • Why timing, not surgery, determines patient survival

      Michael Karch, MD | Conditions
    • Why health care leaders fail at execution—and how to fix it

      Dave Cummings, RN | Policy
    • How digital tools are reshaping the doctor-patient relationship

      Vineet Vishwanath | Tech
  • Past 6 Months

    • Forced voicemail and diagnosis codes are endangering patient access to medications

      Arthur Lazarus, MD, MBA | Meds
    • Why are medical students turning away from primary care? [PODCAST]

      The Podcast by KevinMD | Podcast
    • How President Biden’s cognitive health shapes political and legal trust

      Muhamad Aly Rifai, MD | Conditions
    • The One Big Beautiful Bill and the fragile heart of rural health care

      Holland Haynie, MD | Policy
    • Why “do no harm” might be harming modern medicine

      Sabooh S. Mubbashar, MD | Physician
    • The hidden health risks in the One Big Beautiful Bill Act

      Trevor Lyford, MPH | Policy
  • Recent Posts

    • Why point-of-care ultrasound belongs in every emergency department triage [PODCAST]

      The Podcast by KevinMD | Podcast
    • Why PSA levels alone shouldn’t define your prostate cancer risk

      Martina Ambardjieva, MD, PhD | Conditions
    • How to handle chronically late patients in your medical practice

      Neil Baum, MD | Physician
    • Reframing chronic pain and dignity: What a pain clinic teaches us about MAiD and chronic suffering

      Olumuyiwa Bamgbade, MD | Conditions
    • How early meetings and after-hours events penalize physician-mothers

      Samira Jeimy, MD, PhD and Menaka Pai, MD | Physician
    • Why medicine must evolve to support modern physicians

      Ryan Nadelson, MD | Physician

MedPage Today Professional

An Everyday Health Property Medpage Today
  • Terms of Use | Disclaimer
  • Privacy Policy
  • DMCA Policy
All Content © KevinMD, LLC
Site by Outthink Group

Leave a Comment

Comments are moderated before they are published. Please read the comment policy.

Loading Comments...