Evaluation of Medical Data collection and analyzation methods

Review of a few  of the common data collection and evaluation methods:
Claims/ Billing data – self explanatory - collection of data based on the billing and claims processing system – focus on ICD and CPT codes
Advantages: utilizes a well known structured and standardized data format with well parameters ( ICD, CPT and HCPCS code)
CHALLENGES: Limited by what is paid for. Limited to evaluation only on items included in billing ( ICD,  CPT and HCPCS codes). Many studies have demonstrated that claims data are very limited or inaccurate in assessing clinical outcomes and, increasingly, payers are focusing on clinical outcomes. Claims data loses data that may be relative to give deeper information on the factors and progression of a diagnosis.  Frequently treatments that are utilized may not have a CPT or HCPCS code- OTC treatments, home therapies and many other treatments utilized – fall through the gaps using this method.
Claims only analyzes diagnoses - no related symptoms.  Frequently, we need more granular details related to the diagnosis to analyze the progression of a pathology. Ie dimensions, presence of drainage, type of necrotic tissue for diabetic ulcers. Similarly, we need more granular details to track efficacy of our treatments.  CPT and/or HCPCS codes only give high level details on the treatment that was used. Unfortunately no granular details are considered that can affect the success of the treatment ie strength of a drug, sizes of screws, plates and other internal fixators and many others – safety and efficacy on these treatment specifics  are not considered using this data collection method.

Longitudinal patient data - repeated observations of the same variables over long periods of time- observing long term progression of pathologies and long term effects of treatments.  These are very good at showing long term effects of disease across sections of society and sharing similar particular demographics.
CHALLENGES: Tracking patients over extended periods of time is challenging – a number of our  patients in the study can be lost to follow up – so long term tracking is difficult.
Time - require a sufficient time frame to demonstrate economic impact. This may be effective for high-mortality and high-cost categories but may not be appropriate for chronic illnesses – anything longer term is more difficult to track long term/ chronic illnesses difficult, if not impossible to monitor
 
Please share your experiences with differences seen in product performance, advantages and side effects between claims billing an longitudinal studies. Your thoughts and experiences on this method of data collection and evaluation are extremely valuable to your colleagues in the Medical Data Community