Leveraging Clinical Data Through Diagnosis Codes as a Successful Oncology Practice
Strategizing patients by determining which patients and what volume of patients will require the most resources is essential to positioning your practice for a successful future. Any business model would identify the needs of a population before defining what services should be brought to a given area, and medical practices should follow suit.
The linchpin to this critical step lies in the quality of your data. The adage, “garbage in, garbage out” still holds true. Decades of a minimalist approach to assigning diagnosis codes leave the typical medical practice somewhere on the spectrum between incomplete data and inaccurate data, neither of which serves the end goal well.
The role of a diagnosis code was rarely seen as much more than something that payers require to get paid, or something used in the setting of a clinical registry. As a result, running a diagnosis code list today from the average billing system would provide a fragmented story at best.
Physicians can often guess the clinical demographics of their patients, but their focus may still be on the primary medical condition, and not on the total picture. For example, how many patients in your practice are dependent on wheelchairs or require supplemental oxygen? How many patients do not have a caregiver within their primary residence to assist them with adhering to their treatment plan?
Even accurately identifying the number of patients with lung cancer who continue to smoke is challenging for most practices. The long-standing mantra of coding applies here—if it isn’t documented, it never happened. In this context, “documented” means that the condition or the social and/or economic risk factors have been assigned a diagnosis code that can be easily mined.
Clinical registries can be helpful in quantifying data for certain patient populations, but they also lack the breadth and the depth of the information that is needed to successfully leverage patient data. It is important to fully implement the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) diagnosis codes to secure the linchpin.
How to Leverage Data Using Diagnosis Codes
Little progress can be made by working backwards. Start today, and follow the workflow of the clinical data. Which ICD-10 diagnosis codes within the system are included in the insurance claim form? Those diagnosis codes on “the problem list”? Those on the patient’s medical history list? Only those in the assessment and treatment plan for a specific encounter? Can you pull a diagnosis list from these fields in your electronic medical records or only from the billing system? What data are available (ie, social history information) but are not coded?
Review the first frequency report of diagnosis codes and isolate the unspecified codes that were reported unreasonably (ie, when it would have otherwise been reasonable to expect the details for specificity). For example, the laterality of lung cancer, the stage of chronic kidney disease, whether the patient has insulin-dependent diabetes, or whether the neuropathy is attributed to chemotherapy, are cases where a specified diagnosis code is warranted.
To identify which ICD-10 diagnosis codes may be appropriate or relevant but are missing from the diagnosis frequency report, determine the types of conditions—clinical or socioeconomic—that will affect your patient’s ability to adhere to their treatment plan or your patient’s overall prognosis. Next, identify specific points within the data workflow where this information can be coded or captured. Finally, provide your team with the necessary resources to assign the appropriate diagnosis codes, and rerun the diagnosis frequency report within 30 to 60 days.
Mining your own data is a necessary step before enlisting the assistance of an external vendor. The diagnosis frequency reports must be able to run on a routine basis without drawing on the expertise of a third-party software company. It is important to be able to organize the data in such a way that they can be leveraged to answer numerous questions, such as:
- The number of physician visits per a specific diagnosis during an exact period of time
- The top 25 diagnosis codes for each provider that were submitted for every level of service
- The number of physician visits that were submitted with only 1 diagnosis code (unusual in oncology, unless the physician visit is 100% surveillance-related).
It is always interesting to run a diagnosis frequency report by provider; many practices have an individual who feels that he or she has the most complex patient cases—check the numbers. Being able to accurately describe your patients through data will mean the difference between successfully navigating the new world of the Medicare Access and CHIP Reauthorization Act and being lost in the sea of acronyms.