*EXCLUSIVE* Flawed federal reports used for health policy and funding decisions

 

bw-tag-4

wp-cdc-reportThis week, several national news organizations ran a story on a new CDC study which showed rural Americans were more likely to die from the top 5 causes of death than their urban counterparts. Per the study, “the five leading causes of death in the United States during 1999–2014 were heart disease, cancer, unintentional injury, chronic lower respiratory disease, and stroke”. While the report results shocked, the data upon which the conclusion is based is deeply flawed, containing a statistical error–an error which my investigation found to be contained in other far more critical HHS reports provided by mandate to the President and Congress.

In short, the error result is a significant over-inflation of the mortality burden of heart disease in America while under-reporting the impact of venous thromboembolism.  To understand the life-and-death implications of how one tiny error can cascade throughout the health system without question, first let’s clearly define what we mean by these two medical terms, both in lay terms and in statistical analysis.

What is considered heart disease?

When you hear the term ‘heart disease’, what do you think of?  Heart attack and cardiac arrest, right?   That’s also what our health agencies educate the public.  From the NIH’s National Heart, Lung, and Blood Institute:

“Coronary heart disease—often simply called heart disease—is the main form of heart disease. It is a disorder of the blood vessels of the heart that can lead to heart attack. A heart attack happens when an artery becomes blocked, preventing oxygen and nutrients from getting to the heart. Heart disease is one of several cardiovascular diseases, which are diseases of the heart and blood vessel system. Other cardiovascular diseases include stroke, high blood pressure, angina (chest pain), and rheumatic heart disease.”

Clinical researchers define heart disease more precisely using the International Classification of Diseases, Tenth Revision (ICD-10), a system used by healthcare providers to code all diagnoses and procedures.  All Health Insurance Portability and Accountability Act-covered entities are mandated by the U.S Department of Health and Human Services to use this classification system. It’s primary application is for billing and reimbursement purposes, but the data collected also provides epidemiological researchers a way to look for the incidence and distribution of specific medical conditions.  Here’s how the CDC researchers in the referenced study used ICD-10 codes to define ‘heart disease’:
cdc-def-1

Heart disease here is defined as ICD-10 codes: I00-I09, I11, I13, and I20–I51.

The problem with this definition is that these ICD-10 codes capture far more than merely heart attack and cardiac arrest.   For example, this range includes code I26, the code for pulmonary embolism.

icd-10-peA pulmonary embolism (PE) is a blood clot in the lung.  Per the NIH National Heart, Lung and Blood Institute:

“PE most often is a complication of a condition called deep vein thrombosis (DVT). In DVT, blood clots form in the deep veins of the body—most often in the legs. These clots can break free, travel through the bloodstream to the lungs, and block an artery.”

“If a blood clot is large, or if there are many clots, PE can cause death.”

While a pulmonary embolism is a blood clot which lodges in a pulmonary artery of the lung to block blood flow, it’s etiology is venous.  “A PE occurs when a blood clot breaks off from a DVT and travels through the blood stream, traversing the right atrium and right ventricle, and lodging in the lung.” Why is this distinction important?  Because….

All blood clots are not equal

Clots in arteries and clots in veins have different risk factors, treatments and risks for recurrence. They are therefore prevented in different ways.

  • Blood clots which form in arteries lead to: stroke, TIA (mini-stroke), heart attack (myocardial infarction), and peripheral arterial clots.
  • Blood clots which form in veins lead to: venous thromboembolism (VTE) which is a term used for both pulmonary embolism-PE (lung clot) and deep vein thrombosis-DVT (most commonly leg).  VTE =  DVT + PE.

Why this matters:  follow the money

In practical terms, if you are a public health official and you want to design a program to prevent blood clots, you’ll need a very different educational tact for tackling heart attack and stroke (arterial clots) than you will deep vein thrombosis and pulmonary embolism (venous clots).  Different causes = different prevention strategies.

For a heart attack prevention program, you’d focus your educational outreach on risk factors such as cholesterol, blood pressure, obesity, and smoking.  For VTE prevention, you’d focus education on risk factors such as hospitalization, cancer, pregnancy, oral contraceptive use and travel.   Very different messages. Very different audiences to educate.

cdc-heart

CDC Division of Heart Disease & Stroke Prevention

Even though the CDC includes pulmonary embolism in the ICD-10 codes it pulls to define ‘heart disease’, in policy practice, PE is excluded from heart disease prevention efforts. The CDC has an entire division dedicated to heart disease and the prevention of heart attack and stroke–the Division for Heart Disease and Stroke Prevention.  Absolutely no where will you find a single mention of pulmonary embolism on the Divisions’s website–no prevention information, no programs in place, no resources allocated. 

This exclusion simply reflects how most non-researchers think of what constitutes ‘heart disease’–it is heart attack and stroke, not deep vein thrombosis or pulmonary embolism.

Yet because researchers do include pulmonary embolism in their statistical definitions of ‘heart disease’, their data over-inflates both the incidence and burden of true heart disease (heart attack and stroke) while downplaying the significance of PE and it’s underlying cause, DVT.   The data makes heart disease look much bigger than it is which translates into  a stronger policy  case for tax-funded  programs aimed at heart disease…which in practice are  programs exclusively targeting heart attack and stroke.

A widespread error in numerous reports used by the President, Congress, reporters and the public

cdc-health-reportThe  flawed definition of ‘heart disease’ is not limited to this week’s small CDC report.  The same definition appears in the 461 page benchmark report, Health, United States, 2015, “submitted by the Secretary of the Department of Health and Human Services to the President and the Congress of the United States in compliance with Section 308 of the Public Health Service Act”.  This mandated report is compiled annually by the CDC to present national trends in health statistics.  It’s data, provided by the CDC Center for Health Statistics, is often used to make the cases for many critical policy and funding decisions.

This report, and the companion monthly vital statistics data update reports, do not present all causes of death, but rather only 113 select causes of death.  VTE is not one of them, but heart disease is.  Unfortunately,  heart disease is also wrongly defined in these reports by including the ICD-10 codes for pulmonary embolism.

cdc-monthly

source: CDC National Vital Statistics Report

To sum:  The key reports provided to the President and Congress for the purpose of making critical health policy and funding decisions are flawed.  It overstates the burden of heart disease and under-reports deadly, preventable pulmonary embolism. 

What are the true top 5 causes of death?

Because pulmonary embolism is wrapped up in the statistical definition of ‘heart disease’, it does not get it’s own attribution as a ’cause of death’. If VTE were pulled out and reported separately, where might it rank?

If VTE were included, it would be the 3rd top cause of death

Table Source:  CDC https://www.cdc.gov/nchs/fastats/leading-causes-of-death.htm

The CDC estimates upwards of 900,000 new VTE cases each year, with 300,000 deaths from pulmonary embolism.  This would place VTE among the top 3 causes of death…more than AIDS, breast cancer and car accidents combined. VTE deaths in America even exceed (by 26 times more!) those due to firearm homicides which we hear more about, which the CDC places at 11,208.  Even opiods–which just weeks ago in December 2016 both President Obama and CDC Director Thomas Freiden called ‘an epidemic’–kills far fewer persons at 52,000 Americans annually.

Another implication:  if you take away the pulmonary embolism cases from the ‘heart disease’ category, then cancer would move up the list as a cause of death, likely making it the true #1 cause of death among Americans rather than heart disease.

Implication:  A public health threat goes ignored

That heart disease figures are over-stated might not be cause for alarm if VTE was being addressed in other meaningful ways.  It isn’t.

The CDC’s FY 2017 Congressional budget justification includes $87 million by name for two other thrombotic conditions, heart disease and stroke,  there is no line-item specifically for VTE which by the CDC’s own mortality figures, kills more Americans annually than stroke

cdc-fy2017-blood-disorders

CDC’s FY 2017 Congressional budget justification

While there is no line-item funding allocation for VTE as you find with other conditions and it’s not included in the mission of the Division for Heart Disease and Stroke Prevention, that is not to say there is not some effort to address VTE by the CDC within its budgetary constraints.   Found within the CDC National Center on Birth Defects and Developmental Disabilities (NCBDDD) is a $4.5 million line-item titled “Public Health Approach to Blood Disorders”.  It is from a portion of these funds and within this division that current CDC efforts to address VTE arise. Past efforts have included creating a Flickr album of patient stories and hosting a webinar for health care professionals. As a VTE survivor and thrombosis educator, I was pleased to contribute to both these educational efforts.  However, despite good intentions to do the most possible with the available resources, we must admit that these are relatively low-impact, virtual based efforts when held in comparison to the more interventional-type methodologies employed to address other preventable medical conditions. For comparison, take a look at the national, hands-on network of programs the CDC has implemented for stroke.

 Impact on public perceptions and lives

A consequence of the current flawed methodology is that it is difficult to advocate for what isn’t reported correctly. What is reported widely as the ‘top causes of death’ influences public perceptions and more importantly, it impacts where policymakers choose to focus limited public health funding.

If we want to call VTE ‘heart disease’, fine, but then our public health agencies whose mission is to tackle heart disease should be utilizing their funds to develop VTE prevention programs. But to present a false picture to the public (and funding lawmakers) of the burdens of both heart disease and venous thromboembolism…well, that’s just plain wrong and people are dying as a result.

How?  Because funds are not being invested on conditions where they are truly most needed. Most VTE are preventable. Nearly a third of VTE result in death. The Agency for Healthcare Research and Quality (AHRQ) calls VTE “the most common preventable cause of hospital death”.  There are effective prevention, diagnostic and treatment strategies. Yet, it is hard to argue that VTE is a serious problem which needs funding and dedicated efforts when it doesn’t even show up on on the most commonly utilized and widely distributed lists of causes of death and disability.

Needed action

I do not like to point out a problem without offering a solution.

A 2008 Surgeon General Call to Action report found VTE was “a major public health problem, exacting a significant human and economic toll on the Nation”. The report found gaps existed in the systematic application of clinical knowledge and that the condition itself suffered from low public awareness….both ideal preconditions for a coordinated public health intervention to have impact. But we’re so far off from where we need to be that it is going to take funding attached to a Congressional mandate to create one.

While we would like to ideally think that our public health agencies prioritize those medical conditions which pose the greatest threat to American’s morbidity and mortality, this is not the case. While there are effective strategies to prevent and treat VTE, it is a largely ignored public health concern.

CDC is the agency charged with protecting the public health from preventable conditions. The problem with the current CDC approach to VTE is 2 fold:

  1. VTE is not a blood disorder. It is also not heart disease.  VTE is a cardiovascular event. There are identifiable risk factors for VTE–such as hospitalization, immobility, cancer, long-distance travel, oral contraceptives, pregnancy, advanced age. So it doesn’t fit neatly within the mission scope and skill set of an administrative unit at CDC which primarily addresses conditions acquired at birth, which is where the topic is languishing now. To capitalize on existing expertise and tap economies of scale, it should be addressed by the same administrative unit as the other 2 major thrombosis-related conditions–heart disease and stroke–within the National Center for Chronic Disease Prevention and Health Promotion.
  2. Because VTE is not mentioned by name, with a line-item expressly for it, funds in the ‘Public Health Approach to Blood Disorders’ line can easily be diverted and utilized on non-VTE educational activities, which has been seen in past years. The only way to ensure funds are utilized for VTE is for it to have its own, clearly named line-item allocation.
To clean up our health data  and target VTE, the President and Congress need to:
  1. Establish a task force (as has been done for other conditions)  to review, consolidate and coordinate all US VTE efforts across federal agency stakeholders–CDC, CMS, AHRQ, NIH, FDA, HRSA, Veterans Affairs, Rural Health.  Currently, efforts are siloed reflecting a lack of cross-agency collaboration.
  2. Make a clearly named VTE line-item funding allocation.
    1. Funds should carry the mandate that they can only be utilized for VTE, prohibiting the ability to divert funds towards other conditions.
    2. A clear line of reporting and accountability should be implemented with attached funds to ensure results.
      1. If it is determined that CDC should maintain responsibility for public health prevention of VTE, then a clearly named VTE line-item funding allocation must be made within the same CDC administrative unit that addresses other thrombotic conditions–the National Center for Chronic Disease Prevention and Health Promotion.  If another agency is chosen, VTE should be paired with related thrombotic conditions to take advantage of existing expertise.  If VTE is going to be statistically treated like heart disease, it should get the corresponding resources to go with it.
  3. Establish a work group to review current health data reports and decide upon a consistent methodology for data definitions.  I recognized the coding issue with pulmonary embolism because I know this condition very well, but it calls into question…could there be flaws in the definitions of other medical conditions?   We have become such a data-driven system for decision making, we need to have confidence in the quality of the data.  A thorough review and revision is in order.

Reference

  1. Moy E, Garcia MC, Bastian B, et al. Leading Causes of Death in Nonmetropolitan and Metropolitan Areas — United States, 1999–2014. MMWR Surveill Summ 2017;66(No. SS-1):1–8. DOI: http://dx.doi.org/10.15585/mmwr.ss6601a1.
POST UPDATE:

I corresponded with the CDC study author, Dr. Enest Moy, who was promptly responsive, polite and helpful (emails copied below).  He confirmed that indeed, the codes for pulmonary embolism are included in the statistical definition of heart disease. However  I’m told the fix may be difficult and their data shows that even if PE were pulled out separately it would only account for 7,000-9,000 deaths which is far fewer than reported by other HHS agencies (CDC,AHRQ,CMS)and independent academic researchers funded by HHS (CDC,NIH) to study the epidemiology and incidence of VTE.(Significant CDC and NIH funded work in the field of modeling VTE incidence data has been led by Dr John Heit of the Mayo Clinic, see publication citation) Without digging further to understand the disparity, I cannot say for certain why different agencies have different death figures, but clearly  not all these figures can be correct…the disparity is too wide.  I suspect the difference may be due to how a cause of death is defined between various datasets. Additionally, death certificate data is notoriously inaccurate when it comes to pulmonary embolism.  Again, I cannot say based upon the information at hand, but this makes a stronger case for my recommendation #3:  that a consistent methodology is needed across federal health data reports. If we’re making policy and funding decisions based upon data, we need clean, trustworthy data. At the moment, I don’t have full confidence in the data I’ve seen.  BW

One Comment

The current state of the health care establishment is abysmal. This includes the selection process and training going of future physicians. No easy fix but honesty from the beauracracy would be a good start. Written by a retired ED physician.

Leave a Reply

Name and email address are required. Your email address will not be published.

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

You may use these HTML tags and attributes:

<a href="" title="" rel=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <pre> <q cite=""> <s> <strike> <strong> 

%d bloggers like this: