Summary of Coding Discrepancies
Between DSM-IV-TR and ICD-9-CM

Although all diagnostic codes listed in DSM-IV-TR are legal and valid ICD-9-CM codes, there are still enough small technical differences between the two systems to have justified the publication of “DSM-IV Crosswalk: Guidelines for Coding Mental Health Information”, published by the American Health Information Management Association (AHIMA) in 1999.   One differencwe has to do with the order in which the codes are listed on the medical record (see below).   Other differences have to do with the fact that coders working in hospital settings generally use the ICD-9-CM Alphabetical Index in order to find the diagnostic codes that correspond to diagnoses noted on the medical chart rather than the DSM-IV-TR itself.   For technical reasons, in some cases coders can end up with different diagnostic codes that they would have had they looked the code up in the DSM-IV-TR.  Note that although there are differences, in both cases the submitted codes are valid and will meet HIPAA requirements.  

1) Differences Based on the Order of Listing of Diagnostic Codes

This applies to the following DSM-IV-TR codes:

293.0   

293.81

293.82

293.83

293.84

293.89

294.0

294.10

294.11

In DSM-IV-TR, all of these codes are for mental disorders due to a general medical condition.  In ICD-9-CM, these codes are considered secondary to the general medical condition and thus the general medical condition must always be listed first.  This is the one difference that might technically lead to a “coding violation” in that rules governing the order in which diagnoses are listed is violated.  In DSM-IV-TR, use of any of these codes requires the DSM-IV-TR user to list the causative general medical condition on Axis III.   However, in ICD-9-CM, the general medical condition must always precede the psychiatric condition.  If the diagnoses are listed in the order that they are written using the conventions of the multiaxial evaluation, the general medical condition, which in ICD-9-CM is supposed to ALWAYS precede the 293 or 294 codes, would instead follow it, which is a clear violation of ICD-9-CM rules.  Thus, at a minimum, users must be cautioned to switch the order of presentation of these codes.

2) Other ordering issues

Another ICD-9-CM rule is that the first diagnosis listed is the "condition that occasioned the admission of the patient to the hospital" or was the “focus of the outpatient visit.”  When using the DSM-IV-TR in mental health treatment settings, presumably it is the psychiatric rather than medical condition that was the reason for visit or admission to the hospital, so ordinarily there would not be a problem.  However, psychiatrists practicing in general medical settings like primary care clinics or consultation-liaison settings (i.e., who are treating the psychiatric sequalae of a medical condition should list the medical condition first.  This problem is less likely to be detected by software since the computer would have no way of knowing whether the principal diagnosis is a psychiatric or general medical condition.  (i.e., this differs from the prior situation in which the codes themselves dictate the expected order).

3) Differences Based On 5th digit codes.  

In some circumstances, ICD-9-CM requires the coder to select a 5th digit code for a diagnosis based on chart information.   DSM-IV preserves several such instances of this; most familiar to psychiatrists is the selection of the 5th digit code for Major Depressive Disorder  and Bipolar Disorder to indicate current severity of the mood episode.  It turns out that ICD-9-CM has a number of other such 5th digit choices that are no longer reflected in the DSM-IV-TR because they represent distinctions that are not specifically recognized in the DSM-IV-TR or that have been superceded by more complex distinctions.  DSM-IV-TR deals with this by automatically selecting "0" as the fifth digit.  However, a coder following ICD-9-CM rules is supposedly obligated to choose the appropriate fifth digit code based on the patient's clinical information.   An example of this is seen in the 5th digit codes in ICD-9-CM for alcohol and drug dependence.   The choices are "0=unspecified; 1=continuous, 2=episodic, and 3=remission."  DSM-IV-TR does not care about "continuous vs. episodic" and employs much more complex remission specifiers than just saying "in remission".  Thus, using the DSM-IV codes of 304.00 for alcohol dependence is, strictly speaking, inadequate because it does not take advantage of the 5th digit code when such information (such as the fact that the patient is in remission) exists.  However, since 304.00 is in fact a legal ICD-9-CM code, it should not be rejected in a claim.  The 5th digit code for alcohol dependence has not been used since DSM-III-R; that is, since 1987, and psychiatrists have been submitting claims with 304.00 without known problems  (Of note, this 5th digit was included in DSM-III but dropped from DSM-III-R and DSM-IV).

Categories with 5^th digit codes not used in DSM-IV-TR include: Schizophrenia and related disorders (categories 295.1x, 295.2x, 295.3x, 295.4x, 295.6x, 295.7x, and 295.9x), Pervasive Developmental Disorder (299.0x, 299.1x, 299.8x), and Alcohol/Drug Dependence/Abuse (305.0x, 305.2x, 305.3x, 305.4x, 305.5x, 305.6x, 305.7x, 305.9x).

4) "Poisonings vs. adverse events" 

The diagnostic codes in the 292 section (i.e., DSM-IV-TR codes 292.0, 292.11, 292.12, 292.81, 292.82, 292.83, 292.84, 292.89, and 292.9) are for drug-induced psychiatric symptoms regardless of the context in which the drug is used.  ICD-9-CM makes a distinction between "poisoning" (a broadly defined term for drugs that are "not used according to proper instructions" includes intentional overdose, accidental overdose, combining it with alcohol, etc.) and "adverse reaction" (drugs used as prescribed but that lead to symptoms as side effects).  For cases of "poisoning" ICD-9-CM requires first a poisoning code (from the 900 section of ICD-9-CM), and then the 292 code, and then the "appropriate E code from the Table of Drugs and Chemicals to indicate Accident, Suicide Attempt, or Assault" .  For cases of "adverse reaction", ICD-9-CM requires first the 292 code and then the "appropriate E code from the Table of Drugs and Chemicals to indicate Therapeutic Use."  This was dealt with in part in DSM-IV-TR on pages 879-882 in Appendix G, which contains selected Poisoning codes and E codes for Therapeutic Use.  (E codes for Accident, Suicide Attempt or Assault were not included, in order to minimize confusion).   These codes have not been systematically reviewed to insure that they are up-to-date (i.e., they were probably lasted checked in 1994 and not since then given their infrequent use).

5) Other coding discrepancies. 

There are a number of examples where the AHIMA crosswalk indicates different ICD-9-CM codes from what we have in DSM-IV-TR.   Differences include: Bipolar NOS (DSM-IV-TR: 296.89, ICD-9-CM 296.7), Delusional Disorder (DSM-IV-TR 297.1, ICD-9-CM: 297.9); Attention-deficit/hyperactivity NOS (DSM-IV-TR 314.9, ICD-9-CM: 314.01).  These discrepancies are due to the way the disorders are listed in the ICD-9-CM index.

In other cases, two different ICD-9-CM codes are offered based on distinctions among the ICD-9-CM categories which are not recognized in the DSM-IV: (292.81 is used for both intoxication delirium and withdrawal delirium in DSM-IV-TR, whereas the cross-walk suggests 292.81 for intoxication delirium and 292.0 for withdrawal delirium); (303.00 is used for Alcohol Intoxication in DSM-IV; in ICD-9-CM, the code is "303.0x for Alcohol Intoxication with Alcoholism" or "305.0x for Alcohol Intoxication without alcoholism); (307.42 is for primary insomnia in DSM-IV-TR, while in ICD-9-CM the code is 307.41 for "transient" insomnia and 307.42 for "persistent insomnia")(307.44 is for primary hypersomnia in DSM-IV; in ICD-9-CM the code is 307.43 for "transient hypersomnia" and 307.44 for "persistent hypersomnia"  

In some cases, particularly in the sexual dysfunction section, the ICD-9-CM crosswalk is simply incorrect, suggesting that 302.7 and 302.8 be assigned for sexual dysfunctions due to general medical conditions when it is clear from the ICD-9-CM tabular that these condition would be excluded (i.e., 302.7 and 302.8 are for ‘non-organic” sexual dysfunctions).  This "error" actually reflects problems in the ICD-9-CM index, which appears to incorrectly lead coders to the wrong code (at least insofar as the AHIMA person who did the cross-walk interprets the index entries).  These errors apply to the DSM-IV-TR codes 608.89 and 625.8. 

6) Coding of Abuse and Neglect.  

These codes differ in several situations because the ICD-9-CM adopted a very complicated system for coding abuse and neglect, with different codes applied based on the type of neglect/abuse and whether the code applies to the victim or perpetrator.  A table is needed to explain these differences.