ICD-10-CM cannot save the personal health record

In a Boston Globe article, several so-called experts obtusely suggest that the switch to ICD-10-CM will improve the quality of data in personal health records.

First, we must briefly say what is a personal health record (PHR). Then, we will recap the Globe story. Finally, we will illustrate that the use of ICD-10-CM in place of ICD-9-CM could not have helped the gentleman in the story.

Wikipedia defines a personal health record as ...a health record that is initiated and maintained by an individual. This definition does not account for the recent trend of companies like Google and Microsoft setting up personal health records, whereby health care providers and payers also contribute data to alleviate the amount of data entry required by the person.

The Boston Globe article recounts the story of Dave deBronkart, who set up a PHR with Google. Google helped transfer claims data into his PHR from a Beth Israel Deaconess Medical Center.

Mr. deBronkart was subsequently alarmed to see a diagnosis in his PHR of spread of his cancer to his brain or spine. You see, Mr. deBronkart has a history of kidney cancer. But it had previously spread to his skull, not his brain or spine. But there is no ICD-9-CM code for spread of cancer to the skull, so the experts quoted in the article understandably postulate that medical records coders used instead codes for spread to brain and/or spine.

But two experts, Drs. John Halamka and Roni Zeiger, then go on to claim that ...the records will improve as more precise coding language is adopted in the coming years. The article does not mention the particular coding system they had in mind, but since the data in question was claims data, it is hard to imagine otherwise.

So, could the switch to ICD-10-CM have prevented the unnecessary fright experienced by Mr. deBronkart?

NO.

ICD-10-CM has the C64 family of codes for malignant neoplasm of kidney (his primary cancer that subsequently spread to his skull), including C64.0 (right kidney), C64.1 (left kidney), and C64.9 (unspecified kidney).

It also has C79.31 - Secondary malignant neoplasm of brain and C79.51 - Secondary malignant neoplasm of bone.

But ICD-10-CM has no code for secondary malignant neoplasm of the skull.

You cannot use C41.0 - Malignant neoplasm of bones of skull and face, because that code must be used only for cancers of bone that arise in the skull and face, not for any cancer that spreads to the skull or face from somewhere else.

And that's it. There are no other even-close-to-relevant codes.

So much for any benefit to ICD-10-CM to help Mr. deBronkart.

Final rule for ICD-10-CM survives Obama administration review

HHS' final rule mandating ICD-10-CM has passed muster with the Obama administration. The rules will proceed unaltered, and thus the final compliance date for ICD-10-CM is Oct 1, 2013.

Hospitals, physicians, clinical laboratories, health plans, the federal government itself, state governments, nursing homes, and more will all now spend an estimated $1 billion to upgrade from a bad disease coding system to a slightly less bad, but unnecessarily more complicated, one.

Final ICD-10-CM rule likely on hold for review

The Obama administration has held for review all rules that either have not been published or have not yet taken legal effect. The latter condition applies to the final ICD-10-CM rule, because it does not take effect till March 17.

It's a long shot, but perhaps the review will lead the Obama administration to realize what a mistake the ICD-10-CM switch is.

The 318 ICD-10-CM codes for diabetes mellitus

In a previous post, we pointed out that despite the fact that there are very few known subtypes of diabetes mellitus, ICD-10-CM has approximately 290 codes for diabetes mellitus, not counting gestational diabetes mellitus.

In the 2009 release of ICD-10-CM, we count a total of 318 codes for diabetes mellitus, including gestational diabetes mellitus. The reason for the large number of codes is that ICD-10-CM combines multiple disease classes into a single code.

For example, the ICD-10-CM code E11.621 Type 2 diabetes mellitus with foot ulcer, contains two disease classes: diabetes mellitus and foot ulcer. For sure, this code implicitly means that the former caused the latter (note that this causal relationship is inaccessible to the computer), but that augments our point. Nothing is its own cause and thus these two diseases are distinct.

We provide here on Google docs the 318 codes and their text strings, in a spreadsheet format that anyone can at least copy-and-paste into their own spreadsheet or database table. An easy way to demonstrate the needless complexity caused by combination codes.

Myth: SNOMED CT has more disease codes than ICD-10-CM

Because SNOMED CT is a reference terminology, and ICD-10-CM a disease classification, one might think that SNOMED CT would have more disease codes because it reaches a higher level of diagnostic precision (what the ICD-10-CM proponents ambiguously refer to as "specificity") than ICD-10-CM.

One would be wrong, however. We already busted this myth in a previous post, but we give it its own post to highlight the absurdity that is ICD-10-CM.

Per the final rule (warning: pdf) to adopt ICD-10-CM, ICD-10-CM has approximately 68,000 codes. SNOMED CT (the July, 2008 version), by contrast, has 63,731 active disease codes.

ICD-10-CM therefore has approximately 7% MORE disease codes than SNOMED CT. Assuming of course, that ICD-10-CM contains only codes for diseases, which it doesn't. It has codes for lots of other things, like symptoms of disease. If there were any way to count automatically how many ICD-10-CM codes represented diseases as opposed to something else, it would be possible to do an actual apples-to-apples comparison.

But, since ICD-10-CM says it classifies diseases (and not other things) and gives no way to infer automatically (i.e., by computer) whether it classifies other things than disease, we feel justified in making this comparison. It highlights another absurdity of ICD-10-CM: it isn't (entirely) what it says it is.

It's final: ICD-10-CM by Oct 1, 2013

The Department of Health and Human Services issued today a final rule (warning: pdf) mandating the adoption of ICD-10-CM as a code set under the Health Insurance Portability and Accountability Act (HIPAA). It pushed back the deadline from Oct 1, 2011 (from its proposed rule last August) to Oct 1, 2013.

At approximately the same time, the National Center for Health Statistics released a new, 2009 version of ICD-10-CM that is available here. Instead of the 23MB, 2,392 page PDF file of the 2007 format, we now have an 8.8MB, 2,369 page PDF file. A trimming of 1% on the page count, and a shrinking of over 50% in file size.

The health care industry now has a little more than 4.5 years to find every usage of ICD-9-CM codes in all of its systems, and upgrade and test them to use ICD-10-CM. All the effort spent on that, will not be spent on adopting electronic medical records, devising and participating in pay for performance programs, improving patient safety, automating the reporting of notifiable diseases, chronic disease management, quality initiatives, adopting other information technology standards for true interoperability, and the list goes on.

ICD-10-CM fails every basic requirement demanded of modern technology, terminology, and ontology, and yet it--and previously ICD-9-CM which also fails to meet these requirements--are the only code sets the government has mandated the industry adopt en masse. We suppose it's not surprising coming from a government bureaucracy. But it still is senseless.

Myth: It is practical to assign ICD-10-CM codes manually

The proposed rule to mandate the switch to ICD-10-CM states:

It would be impractical to attempt to manually assign SNOMED–CT codes. The number of terms and level of detail in a reference of clinical terminology such as SNOMED CT cannot be effectively managed without automation,...

By implication, then, it would be practical to assign ICD-10-CM codes manually. Otherwise this supposed disadvantage of SNOMED-CT would not be a factor in HHS' decision to reject SNOMED-CT.

Let us examine this claim further.

ICD-10-CM, by all accounts we have seen--including the proposed rule itself, contains approximately 68,000 codes.

First, we think the very notion that the human brain can cope with 68,000 codes and reliably and manually assign a few of them correctly to patient visits or hospitalizations has no face validity.

Second, even with the manual assignment of the 13,000 codes of ICD-9-CM, there is and has been tremendous variability and low reliability. The Department of Veterans Affairs (VA) conducted a study that found substantial variability in assignment of ICD-9-CM codes:

Based on this study, OHI concluded that the coding of the primary and secondary diagnoses varied widely. The implications of this variability has to be considered when assessing the validity of health services research, health care program planning, quality assurance, utilization review, and resource allocation for VA Medical Centers based on ICD-9-CM codes or DRG information.

While OHI was not evaluating the coding "error rate" in this study, the coding variability observed in the study was comparable to error rates noted in earlier Institute of Medicine (IOM) studies. We found a 60.6 percent agreement in the primary diagnosis code among the original coders and our expert coder. The IOM studies documented a 65.2 percent agreement on the principal diagnosis code, in 1977, and a 63.4 percent agreement on the principal diagnosis code of the records analyzed in 1980. Thus, in all three studies there was approximately a 2/3's agreement in the coding of the medical record.

Even among the expert coders, there was a 19 percent disagreement on the primary diagnosis code. Since our expert coders were highly qualified, this high rate of disagreement caused OHI to question the reliability of the selection of the primary diagnosis and, thus, the accuracy of coded information.

A study of ICD-9-CM coding in psychiatry concluded:

The question was addressed how well mental health professionals were able to translate diagnostic formulations into ICD-9-CM codes. This was done with three coder groups and under two conditions. It was found that there was insufficient interrater agreement on the ICD-codes in all groups and conditions. This finding then was related to the inadequacies of the ICD-system itself. It was concluded that current mental health statistics that are based on the ICD-9-CM coding system are without scientific value.

A study of ICD-9-CM coding in intensive care concluded:

In a multicenter database designed primarily for epidemiological and cohort studies in ICU patients, the coding of medical diagnoses varied between different observers. This could limit the interpretation and validity of research and epidemiological programs using diagnoses as inclusion criteria.

Since other nations have already switched to ICD-10 or their own national variant of it (none of which has even half as many as 68,000 codes), what has their experience been with ICD-10? Better coding? No.

One study of the reliability of coding with ICD-10 concluded:

The refinement of the ICD-10 accompanied by innumerous coding rules has established a complex environment that leads to significant uncertainties even for experts. Use of coded data for quality management, health care financing, and health care policy requires a remarkable simplification of ICD-10 to receive a valid image of health care reality.

A study from Canada even compared the quality of coding between ICD-9 and ICD-10 and concluded:

The implementation of ICD-10 coding has not significantly improved the quality of administrative data relative to ICD-9-CM.

So then, manual assignment of ~13,000 ICD-9-CM codes in the U.S. and elsewhere, and the manual assignment of ~13,000-30,000 ICD-10 codes (depending on national variant), have not been "effectively managed".

It brings to mind the old adage, those who live in glass houses should not throw stones.

So what of SNOMED-CT? How many disease codes are we looking at?

The July, 2008 version of SNOMED-CT, by contrast, has 63,731 active disease concepts. ^[1]

SNOMED-CT, therefore, actually has fewer disease codes than ICD-10-CM! It is hard to imagine that manual assignment of SNOMED-CT disease codes could be managed any less effectively than manual assignment of ICD-10-CM disease codes.^[2]

Myth: Busted.


^[1]Because SNOMED-CT, unlike ICD-10-CM, comes in machine-readable format, these kinds of exact counts are easy to make.

^[2]Note that we are not advocating SNOMED-CT for disease coding. And studies conducted thus far have shown lack of reliability in SNOMED-CT disease coding as well.