In his preview¹ in Health Affairs of the impending data deluge scheduled to begin with the transition to ICD-10 on October 1, 2013, Harris Meyer explains that the international disease-classification system dates back to 1763, when—in an act of altruism toward his fellow physicians—Francois Boissier de Sauvages de Lacroix (1706–1767) published a list of 10 major classes of diseases and 2,400 individual diseases.
When the adoption date rolls around in 2013, it is unlikely to be greeted by radiologists—or their coders and billing companies—as an act of altruism. The number of diagnostic codes will jump from 16,000 to 69,000; the number of procedure codes that can be performed on an inpatient basis will jump from 3,800 to 72,000, affecting every aspect of clinical and business operations.
The first International List of Causes of Disease was jointly adopted in 1898 by Canada, the United States, and Mexico, while other countries developed their own lists, some with morbidity listings. The modern ICD got its start in 1948, when the World Health Organization (WHO) took over responsibility for disease reporting and published its first International Statistical Classification of Diseases in 1949: ICD-6. The system provided a framework for tracking and understanding disease trends worldwide and was updated every 10 years.
In 1976, WHO issued ICD-9, the version still in use in the United States (though other countries began adopting ICD-10 as early as the 1990s). Many countries, including the United States, developed their own versions, raising concerns that the system will be less useful in tracking world health trends, moving forward. The US version is by far the most complex, with 69,000 separate codes; Australia’s has 16,000, and Germany’s has 13,000.
The new level of detail contained in ICD-10, however, has the potential to confer many benefits. For instance, ICD-9, Meyer writes, lacked enough granularity to support the move to value-based purchasing, which requires a more precise accounting of appropriateness and intensity. ICD-10 also has the potential to aid fraud enforcement by improving the ability to spot mismatches between diagnoses and procedures. The DHHS believes that ICD-10 will improve efficiency by reducing claims rejections by insurers and by facilitating the identification of patients with chronic conditions.
Beyond the changes required of physicians, coders, and billing personnel in adapting to a new disease classification that has more than a fourfold increase in diagnostic codes and an 18-fold increase in procedure codes, the new system features structural changes that will require big adjustments on the part of providers and payors, Meyer forewarns readers.
First, ICD-10 makes procedure codes separate entities, called ICD-10-PCS. Second, ICD-10 codes are not only alphanumeric, but also several digits longer: three to seven digits instead of three to five for diagnostic codes and seven digits instead of three for procedure codes. Third, there is a very low rate of matching between ICD-9 and ICD-10 codes for similar procedures.
CMS has published general equivalency maps linking similar codes in the two versions, but the Workgroup for Electronic Data Exchange recommends that health-care providers create their own (more precise) crosswalks between ICD-9 and ICD-10. This is particularly important due to the fact that for at least one year after Medicare’s October 1, 2013, go-live date, providers will have to continue to process ICD-9 codes from prior to that date.
In addition, even though providers will be required to submit all payment claims in ICD-10 format, private payors are not required to make the transition, and a significant lag in the adoption of ICD-10 by some private payors is expected. “If providers aren’t careful, those differences could have major reimbursement consequences, including underpayments or overpayments of several times the correct amount, according to CMS,” Meyer writes.
ICD-10, however, captures a far greater level of detail than its predecessor. With ICD-10-PCS procedure codes, for instance, one can differentiate body parts, surgical approaches, devices used, resource consumption, and outcomes. Angioplasty has just one code in ICD-9, but 1,196 in ICD-10, enabling coders to specify the precise location of the blockage and the instruments used.
Note that transitioning to ICD-10 will require more than software upgrades; providers will be challenged