ICD-10: How to assess compliance and perform audits now that it is here

Danette L. Slevinski, JD, MPA, CHC, CHPC, CHRC, CIPM

Prepared for the American Health Lawyers Association’s Fraud and Compliance Forum held September 27-29, 2015[1]

I.  introduction

The ICD-9-CM code set was updated annually since its implementation in 1979 [2]. There were limitations to this code set. Given the limited number of characters, the Tabular list was running out of numbers in several disease categories.[3] Within disease types, the available codes did not describe, with specificity, certain diseases or conditions when in practice, additional specificity such as disease presentation, type or onset was widely used in clinical practice.[4] When ICD-9-CM diagnosis and procedure codes were used for research, public health, epidemiology or comparison purposes, the lack of specificity also limited the conclusions that could be drawn from the data and the value of statistics compiled from the data. Additional information was available in the medical record than what the code could describe. In the international healthcare arena there was a desire to analyze patient outcomes to determine the quality of the care provided, develop payment systems that required high quality and low cost care, and track public health epidemics.[5]

3

After several years of development, in May of 1990, ICD-10 was endorsed by the World Health Assembly and beginning in 1994, the World Health Organization began using ICD-10.[6] The National Center of Health Statistics (NCHS), the entity in the United States responsible for ICD, worked on the ICD-10 development and implementation process by releasing the code set for public comment, and reviewing feedback from organizations such as the American Health Information Management Association (AHIMA) who tested the code set.[7] In December of 2012, the NCHS released its last update of ICD-10 after which the Unites States Department of Health and Human Services (HHS) published a rule requiring the implementation of ICD-10 in Health Insurance Portability and Accountability Act (HIPAA) “Covered Entities” (CEs) by 2011. After the United States Centers for Medicare and Medicaid Services delayed the implementation of ICD-10 three times—first to October 1, 2013, then to October 1, 2014 and finally to October 1, 2015, [8] CEs are now facing the inevitable transition this fall.[9]

II.  Background

The International Classification of Diseases, Tenth Edition (ICD-10) was developed to replace ICD-9-CM in CEs as of October 1, 2015. ICD-10 consists of ICD-10 CM and ICD-10-PCS.[10] ICD-10-CM Official Guidelines for Coding and Reporting replaces ICD-9-CM diagnosis codes for healthcare services provided on or after October 1, 2105 and for inpatient services for dates of discharge after October 1, 2015.[11] ICD-10-PCS Official Guidelines for Coding and Reporting replace ICD-9-CM procedure codes.[12] Since ICD-10-PCS will only be used for facility reporting of hospital inpatient procedures, Current Procedural Terminology (CPT) and Healthcare Common Procedure Coding System (HCPCS) codes for Medicare Fee-For Service claims will continue to be used for all procedures other than those covered by ICD-10-PCS.[13] Claims submitted after October 1, 2105 must be coded using ICD-10-CM for diagnosis codes, ICD-10-PCS for procedures in the inpatient setting and CPT/HCPCS codes for procedures other than those provided in the inpatient setting.[14]

  1. What is ICD10/CM/PCS?
  2. ICD-10-CM has 68,000 Alpha-Numeric codes divided into 21 Chapters. An ICD-10 code is made up of three to seven characters or digits used to describe a patient’s diagnosis. All ICD-10-CM codes start with a letter, the second digit is always a number, a decimal is used after the third character and many of the ICD-10-CM codes require laterality.

[15]

To code in ICD-10-CM, look up the condition in the Alphabetical Index and then verify the code in the Tabular Index by comparing the information in the medical record to the types and subtypes listed under the condition.[16] Where a seventh character is required, such as to describe whether the patient presented for the initial or subsequent encounter, and within that encounter whether the fracture, for example, was healing routinely, the additional specificity called for in the coding sequence must be provided. [17] If there are questions, the coder must follow-up with the provider and must also ensure that any additional information provided is documented in an addendum if the coder uses it in the code. In practice, software is often used by coders to code instead of books. It is important for coders to know the limitations and potential risks of the software that they use so that they know when to utilize coding books to ensure that the correct family of codes or code has been selected. Here is an example of a diagnosis code in both ICD-9 and ICD-10-CM:

o 274.01 Acute gouty arthropathy (In ICD-9, one of 15 codes for gout)

o M10.071 Idiopathic gout, right ankle and foot (In ICD-10, one of 162 codes for gout)

  1. ICD-10-PCS codes contain seven Alpha-Numeric characters used to make up over 71,000 procedure codes. Each of the seven characters corresponds to a unique category:

[18]

ICD-10-PCS coding utilizes three separate materials: Tables, the Index and the List of Codes. Users can look up the code alphabetically in the Index which displays the specific location of the procedure in the Tables. The List of Codes provides a complete listing of all valid ICD-10-PCS codes with their corresponding description.[19] The coder should use all available operative notes and reports to identify the correct code.[20] Each of the seven characters has up to 34 possible values. To develop the correct ICD-10-PCS code, the coder needs to begin with the first character and work through each of the seven, in order, working through the table for each character to select the correct one. For example, for the first character, the coder should consult Table 1 and select the character for each procedure. Obstetrics, imaging and medical and surgical are three examples of different codes for the first character, also known as the section character.[21]

Although the overall number of ICD-10-CM/PCS codes has increased, clinicians and staff who code will need to familiarize themselves with the types of codes most commonly used in their respective practice. Here is an example of a procedure coded in ICD-9 and ICD-10-PCS:

o 8151 Total hip replacement; 0077 Hip bearing surface, ceramic-on-polyethylene (ICD-9 codes)

o 0SR9049 Replacement of hip joint with Ceramic on Polyethylene synthetic substitute, Cemented, Open approach (ICD-10 code)

  1. What are the key similarities and differences:
  2. Similarities between ICD-9 and ICD-10: Both coding systems have a hierarchical structure and a similar format. In both cases you use alphabetical indexes to look up the procedure or condition and then identify or verify the specific code in the tabular index. Manifestations of disease are delineated by brackets[22] and the coder codes to the highest level of specificity possible to the extent information is provided in the record by the clinician. A cancer diagnosed by needle biopsy may have a less specific diagnosis prior to a core biopsy or lumpectomy during which the neoplasm type may be able to be provided with greater specificity. NOS or Not Otherwise Specified codes appear in both code sets although there is less of a use of NOS in ICD-10, especially in encounters beyond the initial encounter.[23] Whereas the use of NOS in ICD-9 may have adversely impacted reimbursement for cases were more specificity was available in the record but missed by the coder or was known but not documented by the provider, in ICD-10, the code may not be complete and therefore invalid.
  1. How is ICD 10 Different? General differences between ICD-9 and ICD-10 include the reclassification of certain diseases into different categories, chapters or locations in the code set based on insight from 30 more years of medical advancement and injuries being grouped by anatomical site instead of by injury type. There are several new code definitions based on changes in medical literature. V (factors such as historical diseases that impact health status) and E Codes (Poisoning and Injury) are no longer in their own sections but are incorporated throughout the ICD-10 code set [24]

a)  There are seven instead of five digits in ICD-10 and there are two code sets for ICD-10, one for inpatient procedures and the other for diagnoses.

b)  The seventh character represents the encounter episode.

c)  “x” is used as a placeholder in ICD-10. The additional space in ICD-10 allows for expansion and the “x” allows the coder to leave out a character for which there is no relevant code or sufficient information.

d)  Laterality (Left, Right, Bilateral) is added.

e)  Fractures can now be described in more detail at all stages of treatment from the initial encounter through to healing. Location and laterality of the fracture, alignment, fracture pattern as well as in detailed interventions such as cast change must be captured in the code if known. If there are healing delays noted in subsequent encounters they must be captured as well.

f)  Combination Codes: There are now combination codes for conditions and symptoms or representations of those conditions. Poisoning codes are associated with various causes. One ICD-10 code reflects two diagnoses or one diagnosis with a corresponding complication or secondary manifestation.

g)  Excludes Notes: There are two types of Excludes Notes in ICD-10. Excludes 1 signals to the coder that the code should never be used together with the code that the Excludes 1 note falls under. The patient can have the first condition or the second condition, but not both. Excludes 2 notes tell the coder that the condition excluded is not part of the condition described by the code. The patient can have both the condition represented by the initial code as well as the condition described by the Excludes 2 note.

h)  New clinical concepts and expanded codes: concepts such as blood alcohol level and blood type are now uniquely describable by ICD-10 codes. Additionally, conditions such as diabetes, substance abuse and postoperative compilations are expanded to allow for multiple different presentations. There is also a clear distinction between post-procedural and intraoperative complications.[25]

i)  For ICD-10-CM, Instead of NOS “not otherwise specified”, ICD-10 has categories for “unspecified” or “other”/ “other specified”.[26] There are few NOS options in ICD-10 PCS.[27] When documentation does not support a higher level of specificity, which is the exception not the rule in ICD-10, nonspecific codes can still be used.[28] NOS or unspecified can be found in diagnoses like gout and neoplasm, but not for trauma, which has a hard-stop requiring the coder to enter the type of encounter (Initial, Subsequent, or Sequela).

  1. Why the change? ICD-10: Over the last 30 years, there have been medical theory, technological and treatment advances that have outgrown the bandwidth and categorization of ICD-9.[29] With ICD-10, reimbursement can more accurately reflect an appropriate value for specific services rendered and the additional information from the code can support medical necessity. Public health data, disease tracking and surveillance will benefit from the new coding nuances such as the ability to select from additional environmental factors. Data can also be compared internationally with nations now following ICD-10. Quality and patient outcomes analytics will benefit from the ability to distinguish between nuances in disease treatments and resulting outcomes. This could lead to the better development of best practices and improved patient safety. Health care providers and entities such as the government and insurance plans seeking to increase quality and decrease cost can monitor more targeted interventions. The increase in specificity of codes can improve analytics and decrease report generation times since less medical record documentation will need to be reviewed.[30]
  1. What providers should have done to prepare:

In order to be successful in transitioning to ICD-10, healthcare providers, health plans and other CEs needed to not only bring themselves into compliance with the new methodology but also needed to make sure that the CEs to which they provide codes to receive reimbursement for treatment, payment and operations were in compliance and their systems were able to share information. The ICD-10 Quick Start Guide outlines Five Steps Towards Compliance[31]:

Step 1: Make a Plan: this step included reviewing available ICD-10 code books, staffing roles and responsibilities, practice management systems, electronic medical record systems, vendor agreements and billing agreements to determine what systems needed to be updated or purchased and which staff needed to be trained. Careful planning included identifying all stakeholders of, reports containing and other uses of ICD-9 codes including those used for operational analytics, defaulted billing, automated care prompts and flows in the electronic medical record for disease management, quality reports, finance reports, billing defaults, external reporting and external communication to the government, affiliates and vendors.

Step 2: Train Your Staff: This step covered the identification of the staff members who needed to be trained, including clinical staff, affiliated clinicians, billing and coding staff and other members of the work force as well as vendors. Training plans for each target group should have been developed and the most prudent CEs would have begun the process of testing the efficacy of education for documentation and coding accuracy prior to ICD-10 implementation.

Step 3: Update Your Processes: Software, forms, vendors and reports should have been updated as per the plan developed in Step 1. Super bills containing the most common diagnosis codes used by a provider should have been updated from ICD-9 to ICD-10. Infrequently used codes should have been removed and ICD-9 codes replaced with ICD-10 code options using code books or General Equivalence Mappings (GEMSs). [32] Needed software updates should have been identified for EMRs and necessary updates installed. Internal and external report algorithms used for quality, payment, research or other operations that were built, based off of ICD-9 codes needed to be rebuilt, identifying the new relevant ICD-10 codes. Finally, steps needed to be taken to ensure that clinical documentation supported and properly utilized the new breadth and depth of codes available. Special attention should have been given to laterality, fractures, diabetes and the encounter type as well as any new or significantly changed coding families that providers regularly used. Documentation should always be as specific as possible but ICD-10 implementation is a perfect opportunity to have dialogued with providers about the new specificity of ICD-10 and the importance of detailed documentation.