** See new comments below.

Original Post
A common question for me is whether or not dentists qualify as acceptable specialty types for risk adjustment. The answer is that only a subset of dentists – namely, Oral Surgeons (CMS specialty type 19) are acceptable sub-specialists from within the dentistry field.

Oral Surgeons are most typically dentists (DDS or DMD) with advanced training in oral-maxillofacial surgery (OMFS).

Specialists in this field handle all things related to head, skull, neck, jaw, teeth whether it be cleft palate reconstruction, facial reconstruction after trauma, dental implants or cancer-related head and neck surgeries.

So, only dentists who designate themselves in the documentation as providers in Oral Surgery or Oral Maxillofacial Surgery are acceptable dental sub-specialists from which you may obtain ICD-9 codes for risk adjustment.

** Addendum 7/21/2011.
Please see Comment #1 for correction. Thanks mistmi01! Oral Surgeons will map to CMS specialty code 19 (maps to NUCC Taxonomy code 1223S0112X), while standard Dentists must designate themselves as CMS specialty code 99 “Unknown Physician Specialty” when they apply for Medicare privileges. So, all dentists are acceptable sources for risk adjustment. Health plans may use NUCC taxonomy codes to distinguish dentists from oral surgeons. The NUCC code for a standard dentist is 122300000X. Oral Surgeons treat a broader number of conditions that are in the risk adjustment model.

Q: Does a health plan need the facility’s (clinic) permission to add a chronic condition to a claim when the evidence of evaluation and treatment is found in the documentation? Our problem is the very rural areas are coders who do not know anything about risk adjustment and are constantly disagreeing with out findings because they think you ONLY code the reason for the visit no matter what information I give them regarding the importance of coding chronic conditions when addressed/assessed and treated. Any guidance you can offer would be a great help!

A: A health plan does not need a provider’s permission to submit documented ICD-9-CM codes to the RAPS database. Any documented chronic or acute condition that a certified coder could abstract from the proper face-to-face documentation (i.e. signed, dated, credentialed by acceptable specialty type) may be submitted for risk adjustment purposes. All of the chart auditing vendors and any coders working for a plan may abstract and code diagnoses that are properly documented and submit to the plan or submit to CMS on behalf of the plan.

Ideally, these additional codes would be submitted on claims in the first place for ease of processing, but you may submit ICD-9-CM codes to RAPS that have been collected in other means, such as a superbill, chart audit, etc.

The 2012 Announcement mentions an update to the physician specialty type list of acceptable sources of physician risk adjustment data.

These changes are effective with dates of service from January 1, 2010.

Added:
(9) Interventional Pain Management (IPM)
(15) Speech Language Pathologist
(17) Hospice And Palliative Care
(27) Geriatric Psychiatry

Deleted:
(70) Multispecialty Group

PDF document from CSSC Operations website:  Risk Adjustment CMS Specialty Types CSSC

The closest thing I have seen to an official recommendation from CMS on how long providers have to sign or amend records is this guidance from the 2008 Risk Training Guide.  While it mostly addresses claims submissions, I believe one could reasonably infer that 30 days is recommended to have the supporting documentation for that claim complete as well.

“What are the responsibilities of physicians and providers?

Physicians must report the ICD-9-CM diagnosis codes to the highest level of specificity and report these codes accurately. This requires accurate and complete medical record documentation. They are required to alert the MA organization of any erroneous data submitted and to follow the MA organization’s procedures for correcting erroneous data. Finally, they must report claims and encounter information in a timely manner, generally within 30 days of the date of service (or discharge for hospital inpatient facilities).”  (Page 3-19)

Keep in mind that for RADV audits, CMS will allow a provider to sign a CMS-generated attestation that an unsigned document is theirs.  But, only the rendering provider can sign the CMS attestation, partners or other doctors in the practice cannot.

You may or may not be surprised to found out how often providers are generating uncodeable progress notes due to poor word choices for their assessments.  On audits, I still see Diabetes Type 2 described as “DM”, “NIDDM”, or “Diabetes controlled with meds”.  All of these terms fall short in clearly defining the kind of diabetes a patient has.  Education then, is simply a matter of reminding providers to use the phraseology from the most appropriate ICD-9 codes, i.e. the ICD-9 codes that specify “Type 1″ or “Type 2″ diabetes and especially if the diabetes mellitus is controlled or not controlled.

 A second common documentation mistake is to see a diagnosis show up in the Assessment, like “Hypertension” or “Hypercholesterol”, however there is absolutely nothing else in the note to corroborate the diagnosis – no mention of pertinent meds like Norvasc or Lipitor, no recent labs, or no mention of symptoms or review of systems related to the diagnosis.  Such “free floating” diagnoses with nothing underneath to support them might fail a Medicare validation audit.  An exception may be made for some chronic conditions that are mentioned in the patient history and that clearly influence current patient treatment- in these cases, the provider does not have to “prove” in every note that the chronic condition meets the proper diagnostic criteria.

Remember, CMS gives you all the info you need in their monthly Model Output Report to determine which HCCs have been captured in the prior period and have not been yet reported for the current period.  Your IT department can construct a list for you by comparing older MORs to newer ones to see which HCCs are no longer reported.  After mining all of your current ICD-9 claims data and sending in your most current RAPS files – if those HCCs are still missing, it is a great opportunity to prompt your providers to either get that patient back in clinic for followup or to search their prior clinic record for evidence of documentation that would support that missing HCC.  The providers can report those previously unreported ICD-9 codes to the plan with the minimal data set needed for RAPS or can submit another HCFA-1500 provided that the plan has a mechanism to ensure that they can distinguish between new HCFA-1500s that demand payment or merely are “repeat HCFA 1500s” that just carry risk information to retroactively report ICD-9 codes that did not get reported on the original claim.

Diagnoses from problem lists should not be submitted for risk data unless they are supported by a progress note.  The gold standard is that a risk diagnosis must be made each year by an acceptable provider and must be documented each year in a progress note to demonstrate that it is an active condition (or a stable condition that influences medical decision making).  In other words, the patient had to have received services related to that diagnosis each year.  Simply referring to the problem list in a note does not count either.  In other words, you cannot document “Assessment – See Problem List dated 3/30/08″. Each diagnosis must be separately pulled into and supported by the medical progress note.

 EMRs are notorious for importing problem lists from prior visits that may be inaccurate.  Often, erroneous data just gets perpetuated until someone corrects it.

So, providers should use the problem list to remind them to evaluate the current status of all the patient’s conditions, and auditors can use the list to guide their search through progress notes in the chart.

For risk adjustment, it is important to determine the etiology of a pneumonia.  ICD-9 code “486 Pneumonia, Organism Unspecified” carries no risk value, while specified bacterial pneumonias do carry risk value.  It is important to follow published guidelines and clinical judgment when determining the necessity of getting a sputum culture before antibiotic therapy.  The Infectious Diseases Society of America (IDSA) considers it standard of care to get two blood cultures and a deep-cough expectorated pretreatment sputum culture on all patients admitted to the hospital for pneumonia.  Ambulatory patients who are otherwise healthy will probably not benefit from a sputum culture, as they will likely respond to empiric broad spectrum oral antibiotics.   Clinical judgment comes into play when determining the value of a sputum culture in patients who do not quite meet admission criteria, but who are higher risk for failing outpatient management, such as patients with multiple comorbidities, especially baseline COPD, or patients who have multiple antibiotic allergies.

Given the prevailing view that many providers will never become astute clinical coding experts, a plan or medical group should also look to other avenues besides provider education to improve risk reporting.  Task one should be to reduce the number of beneficiaries who skip through a calendar year without a face to face encounter.  Some plans report no encounter rates as high as 10% of their members.   Sure, those patients that manage to skip through a year without being seen are likely fairly healthy, but just finding and coding a couple chronic diseases in a few patients will more than pay for the cost of all their office visits – and even better – yields more opportunities to catch up on preventive care.

Secondly, providers should use the problem list as a prompt to evaluate the current status of and to document their chronic disease states.  Each face to face should be viewed as a chance to document and report these conditions to CMS.

So, plan on developing comprehensive ways to survey your patients’ visits and find mechanisms to get them back to clinic before the end of the calendar year to give providers a chance to evaluate and document their conditions – and more importantly – improve their chronic conditions.

One thing to think about as plans develop their provider education programs is to do what Stephen Covey (7 Habits of Highly Effective People) reminds us to do with Habit # 2: ”Begin with the end in mind.”   In this case, the End is the proper documentation and coding of face-to-face visits that culminates in a solid medical record that will stand up to CMS’ risk adjustment validation audits.  Of course, proper documentation is the goal for many other reasons as well.

The basics include: legible, date of service, contain provider’s signature and credentials, and diagnoses in the assessment that are written clearly so that coders can match them with the correct ICD-9.  Avoidance of “probable”, “rule out”, and “suspected”.  And avoidance of “History of” if the condition is still active.

Also, each ICD-9 diagnosis must be supported by that day’s record for it to be valid to substantiate an HCC code.  That is, a physician cannot write “New onset Diabetes mellitus Type II”, without recording other elements in the note to support that claim, such as mentioning lab data that support this (fasting glucose > 126) or symptoms (thirsty, increased urination).  Any ICD-9 code without underlying symptomatology, history or exam findings, provider-interpreted lab or radiology findings, etc or treatment considerations to support it, will not substantiate an HCC code.

The most likely ICD-9 codes that will not substantiate HCCs are the diagnoses that may have impacted that patient’s care that day, but were not specifically diagnosed, treated or thoughtfully taken into consideration that day.  For example, suppose a patient just has one encounter with me during the year, let’s say he comes in as a new patient with cellulitis and diabetes.  If I only treat his infection and do not ask about his diabetes, and merely list diabetes as a secondary ICD-9 code, I have not done anything to substantiate that diagnosis, that is, I have not written anything in the record that supports or validates diabetes.  However, had I recorded in the history that he told me his last several glucometer readings were about 220 and that he uses an insulin sliding scale at home, then a reasonable coder could abstract diabetes as an active condition supported by the record.

Again, all is not lost, as there will most likely be another date of service dealing with diabetes. But, you may be surprised to learn that in any given year, some plans have no recorded encounters for 10% or more of their beneficiaries.