How Prescription Opiate Overdose Data is Collected– Part 2

Share the News
  • 4
  • 1

By R Carter

Updated September 2019

Since the first draft preparation of this post nearly a year ago, see Part 1 here, much has changed. The CDC’s methods of preparing it’s 2016 Guidelines for Chronic Pain had come under widespread criticism for the disregard of a scientific method in preparing them. At first this was nothing more than an embarrassment, but in the years that followed, do to widespread misinterpretation and the lack of a response from the CDC, it became a political football of sorts as reports of patient deaths and suicides began surfacing. And in nearly every case, the CDC guidelines were cited as the reason for why actions were taken that lead to such deaths. Now, nearly three years later, the agency finally seems to be listening, with two publications in April that take a hard look at the guideline and how they are applied in clinical practice. In his blog, Josh Bloom said it best, The CDC Quietly Admits It Screwed Up Counting Opioids. It’s evident from all of this, that the motivations of the CDC from the beginning were political and not scientific or health related.

Part of the machinery to justify the political motivations, is in how overdose deaths are counted to begin with. This series takes a hard look at how that process works, how a single overdose death can in turn, be counted up to as many as six times, inflating the total count.

February 2018

The CDC’s methods and methodologies are increasingly under fire with regards to it’s reporting on opiate overdose deaths and prescribing guidelines. It’s unscientific methods and disregard for federal guidelines in performing research have more watch dog groups like Washington Legal Foundation taking a swipe at their reports and policy. In Part 1 of this series I looked at the raw sources of data for compiling statistics on prescription opiate overdose deaths. From this we see that 92 percent of the sources are either primary care providers or attending physicians who are required to fill out a death certificate within 72 hours of a person’s death. And we see that general guidelines have no requirements to ensure that a cause of death, listed as an opiate poisoning, be documented by lab tests as proof that toxic amounts were actually the cause of death.

From the CDC Wonder Website, Mortality Data

The mortality data are based on information from all death certificates filed in the fifty states and the District of Columbia. Deaths of nonresidents (e.g. nonresident aliens, nationals living abroad, residents of Puerto Rico, Guam, the Virgin Islands, and other territories of the U.S.) and fetal deaths are excluded. Mortality data from the death certificates are coded by the states and provided to NCHS through the Vital Statistics Cooperative Program or coded by NCHS from copies of the original death certificates provided to NCHS by the State registration offices. For more information, see Technical Appendix from Vital Statistics of United States: 1999
Mortality Data Sources:

The Underlying Cause of Death data are produced by the National Center for Health Statistics (NCHS) at the Centers for Disease Control and Prevention (CDC). Mortality information is collected by state registries and provided to the National Vital Statistics System. Underlying cause of death and demographic descriptors are indicated on the death certificates. Each death certificate contains a single underlying cause of death, up to twenty additional multiple causes, and demographic data. The underlying cause-of-death is defined by the World Health Organization (WHO) as “the disease or injury which initiated the train of events leading directly to death, or the circumstances of the accident or violence which produced the fatal injury.” Underlying cause-of-death is selected from the conditions entered by the physician on the cause of death section of the death certificate. When more than one cause or condition is entered by the physician, the underlying cause is determined by the sequence of conditions on the certificate, provisions of the International Classification of Diseases, and associated selection rules and modifications.

To learn more about the methods and source of these data please reference:

In Part 2 we going to look at how the individual certifying the death certificate uses their knowledge of medicine to make a judgement about what the cause and contributing factors are for a death. I will also look at some of the rules used for coding death certificate cause of death.

Box 32 Part 1

The U.S. Standard Certificate of Death Box 32, Part 1 provides spaces for the certifier to list, in the order of importance, those conditions which lead to the cause of death. They are broke down in sub-parts A, B, C and D and require a code using ICD-10 (International Classification of Disease 10th Revision) codes.


Sub-part A is the immediate cause of death or the final disease or condition resulting in death. Sub-part B-D are a sequential list of conditions or diseases contributing to the cause of death. Based on coding rules, a Sub-part D condition must lead to the cause in Sub-part C and on up the list until the immediate cause is identified. Sub-parts B-D are considered to be “Underlying Causes”, meaning they play an indirect but substantial role in contributing to a persons death, so much so that one can’t occur without the other.

For example a blood infection may attack the cardiovascular system causing a profound and prolonged drop in blood pressure and blood flow. While not being a cause of death, the loss of blood pressure leads to death of cells in the kidney, which produces kidney failure. The kidney failure is ultimately the cause of the death, but the underlying blood infection is a major contributing factor.

The example given above is a critical concept in the discussion about coding a death certificate with codes which list the cause of death as poisoning. Using a similar example as those above we can apply the same logic to a cause of death which involves opiates.

Assume a person’s death is listed as asphyxiation, but the underlying and causes listed in Sub-parts B-D, may be poisoning from Barbiturates, Alcohol and Opiates.

The actual sequences of events were consumption of Opiates, then Barbiturates, then Alcohol, which led to producing a profound level of sedation, not unlike an anesthetic. The body reacts to these as if they were a poison, trying to expel them from the body through vomiting. Because of the profound sedation, the individual is incapable of protecting themselves during this event and the airway becomes blocked by the stomach contents. This produces asphyxiation and death. In this example opiates are not “the cause” of death, but they played a minor role, along with the other drugs.

It is important to note that in more than 95 percent of drug poisoning deaths, there are always multiple substances involved. No one substance is usually the cause. Drugs are synergistic, meaning they compound the effects of other drugs. This compounding phenomenon is a singular cause by itself as no one drug by itself, in any quantities detected, can produce death.

More importantly, to prove a single drug as a cause, you must have toxicology data that both identifies type and quantity of the substance. Further compounding the collection of opiate death data is the fact that in approximately 1 in 5 drug overdose deaths, no specific drug is listed on death certificates. This injects a massive 20% error rate into the data. Also complicating the accuracy of opiate overdoses is that some synthetic opioids, i.e. IMF (Illicitly Manufactured Fentanyl), do not yet have designated codes which can be reported to the CDC. To overcome this shortcoming, state agencies often rely on data from the National Forensic Laboratory Information System which provides information on drug products obtained by law enforcement that tested positive for these Fentanyl analogous, and the National Poison Data System which provides information on poison center calls regarding fentanyl overdose.

It’s for these reasons that when compiling statistics related to prescription opiate overdose deaths, the lines get blurred and ambiguous. It is the inherent ambiguity in these methods which is the Achilles heel in the CDC’s reporting. Strictly speaking though, from a morbidity and mortality point of view, opiates would have been nothing more than “a minor data point” in the contributing cause, because in practice, the entire chain of events is needed to truly understand the role opiates play in overdose deaths.

How One OD Death is Counted as Six

With regards to prescription opiate overdose deaths, report accuracy could be improved by eliminating deaths which include in the causal chain, other substances. It’s a well known fact that in drug overdose deaths, multiple drugs are identified in about 5 out of every 6 deaths. The far majority of OD deaths include two or three substances, but in a small number, as many as six drugs have been identified. For example an OD death which includes hydrocodone, the prescription drug, could include Methamphetamine, Cocaine, Heroin, Fentanyl, Marijuana and Alcohol. Such deaths strictly speaking, are not deaths from prescription opiates, since doctors do not prescribe these other substances. This is just common sense which the CDC methodology doesn’t take into consideration. The slight of hand which is used in tallying the totals, is the fact that hydrocodone is classified as a semi-synthetic opiate, known to be prescribed by doctors. It’s this classification alone that justifies counting the OD death as a prescription opioid death, when in fact, there were five other drugs present and anyone of them could have been the most abundant drug in the victims system. The logic used by the CDC is a unbelievable stretch of imagination, yet this is the basis on which OD deaths are published.

To close this gap and increase accuracy, there have been multiple studies done where a smaller sample of poisoning deaths have complete toxicological studies done to statistically identify specific samplings of prescription vs drugs. Statistical variations from these studies are then applied to death data from mortality records to provide a computed value for prescribed opiate overdose deaths. See Part 3. Although encouraging in its efforts to clarify, even this effort falls short of accurately representing prescription opiates. Those who died from an overdose have still been selected based on a single data point, absent of the other chain of events we know contribute to every death. A good example of this is this report from Texas.

The real problem arises in how the CDC uses this information and reports it. Rather than expressing opiates as a contributing factor in morbidity and mortality, the CDC makes a very black and white assertion that opiates are the cause of death, most often based on this single data point.

By singling out one data point in a sea of other contributing factors, the CDC performs a most grievous disservice to the public. Splitting hairs so to speak, to forward an agenda which is likely motivated by other factors, not the least of which may be politics and ideology. Without an investigation of the process and individuals who operate in this manner, we will never know what their motivations may be.

It’s certainly within the power of the CDC to be more transparent with the public on where the data comes from and how it is used. The question becomes, why don’t they?

Other factors which raise questions about how death certificate data is compiled are based on a system the NCHS uses to capture data from Box 32 Part 1. As I mentioned in Part 1 and here, the certifier filling out the death certificate is only obligated to fill in a blank. What goes into the blank is completely dependent on the qualifications and integrity of the individual. Should that person enter information which is unclear and ambiguous, such as a nondescript poisoning, the NCHS’s Automated Coding System will make an attempt to select codes it believes are most appropriate. From the CDC’s “Coding and classification of causes of death in accordance with the Tenth Revision of the International Classification of Diseases” publication we find this section which describes that process.


While the language and terminology of medicine is quite specific, it’s no guarantee that a word will be interpreted correctly. This is why each diagnosis is required to have a code from the ICD-10 Manual. Between both sources, the system should be able to make an accurate interpretation of the intent of the certifier, most of the time.

What Are The Facts
  • Diagnosis Codes used in Box 32 Part 1 are dependent on the qualifications, intent and integrity of the certifier filling out a death certificate.
  • Because one action or event can lead to another, events leading up to the cause of death are like dominoes.
    • Each is a data point and each subsequent event cannot occur without the preceding event leading to a cause of death.
  • The very nature of the information collected, biology and medicine, makes it problematic to report that information out of the context.
  • With regards to reporting on a single data point:
    • Taking a single data point out of a chain of dependent events is at best misleading and likely irresponsible and unethical.
    • Such reporting is a disservice to the tax payers who pay for this effort and who expect honesty and integrity from those responsible for these duties.
    • When all data points in a chain are required to produce a results, a single data point is not a reliable source of data to form a conclusion leading to prescription opiate overdose deaths. This fact alone disqualifies the single data point conclusion regarding opiate overdose deaths.
    • Reporting a single data point as a singular cause when a dependent chain is required to produce an outcome, is a violation of every known, good and proper procedure for scientific research.
  • Efforts to automate data collection for cause of death, through intelligent software systems, may itself, be a contributing cause for misreporting and misrepresenting facts.
  • The failure of researchers who use death certificate data to divulge how they compensate for the inherent ambiguity in the data, immediately brings into question any conclusions drawn from such research.
  • Good research will identify methods used to improve statistical reliability and prevent contamination of the research and therefore the conclusions drawn.

For the sake of argument. If using a single data point from a chain of dependent events were an appropriate method for identifying problems which justified a massive government regulatory effort to correct, then every data point in the chain would need to be treated with the same and equal weight or importance.

Failure to do so would allow a chain of events to resume from the remaining causes. Under this logic, the same actions and rules applied to opiates, also apply to any contributing drug, including alcohol. Yet we see no efforts to restrict, limit or regulate alcohol on the same level we see with opiates.

For this reason, I continue to believe the war on opiates is and has been an ideological war. Ignoring this logic, refusal to treat all psychoactive substances as poisons, while continuing to push a the singular cause, such as opiates, is another example of extremism and the righteous mindset of an ideological minority.

Preview of Part 3

In Part 3 of this series I will look how the data is weaponized to facilitate the ideological zealots. While I don’t believe the original intent was meant to be an effort that allowed weaponizing it, the inherent shortcomings in both what and how it is collected, allows this to occur.


4 thoughts on “How Prescription Opiate Overdose Data is Collected– Part 2

  1. This discussion brings to my mind another couple things where the data is faulty or at lest murky. Is “accidental” poisoning really accurate? Do they consult with the deceased by Oujia board as to motive? And should accidental be the proper term in cases of overdose? The intent was to get as close to the death experience as possible or they would not be mixing alcohol with benzodiazipines and opioids. Who is to say it wasnt suicide from intractable or cancer pain? If we are going to hold doctor responsible in some of these cases, intent of the patient accidental or deliberate should be an important factor.

    Secondly, in regards to fentanyl, the government and media should share some responsibilities in these deaths. By simply calling it fentanyl and not the more specific type of fentanyl analogues, they are giving abusers the idea that they can survive these drugs, based on their experiences with prescription fentanyl, not an analogue 100’s of times stronger than anything they have experienced before. Granted, drug abusers are not a group that make caution a very high priority and defining these analogues for what they are may even encourage some to seek them out. But there is a population of addicts that would be more discriminating in the drug abuse if they knew they were getting something thousands of times stronger than Grannies Pain Patch. Of course discriminating between prescribed pharmaceuticals and illicit elephant anaesthesia would certainly throw a money wrench in the idea of physician responsibility for the opioid crisis, something the CDC will never allow to be public knowledge. But the least they could do is specify and separate Rx fentanyl from carfentinal and sufentinal to save a few lives…

Comments are closed.