This is lesson 2.2.2, the Elements of Claims Data. In this lesson, we'll discuss what types of information is included in claims data, or administrative data. In claims data, there are five main types of information. The first is membership information. In most claims data, this includes gender, age, and location, whether it's state, zip or even address. The second type of information are the dates and place of service. So we'll be able to tell whether it's inpatient claim, outpatient claim, professional claim and many times within that whether it was in a surgery department In an. If it was an outpatient claim, if it was an MRI clinic, etc. Additionally, there's information about the diagnosis. Does a patient have a diagnosis of obesity? Do they have a diagnosis of heart failure, or congestive heart disease, etc. There's also codes for the procedures or drugs that are administered during an office visit. So this may include the type of surgery that was performed, or if it was a simple wellness visit or preventive check up. If any type of drug or vaccine was administered, this would be listed. And finally, there's often a minimum amount of provider information related to the place of service. It may be possible to identify the provider, whether it's a hospital or physician or facility, outpatient facility and then aggregate claims over that place of service in order to compare places of service or facilities within the claims over time or to each other. Within the claims data there's numerous types of codes that are used to record all this information. Diagnosis codes, Are listed by ICD-9 codes. These are currently five digit codes. The first digit is a number or a letter and all of the remaining elements of that or the other four digits are numeric. However ICD-9s are in the process of being transitioned to ICD-10s. These will be seven digit codes, all alphanumeric and there will be about 68,000. Currently, there's about 13 thousand ICD-9's. The point of the increase in the number of digits and the number of codes is to add more precision to the diagnosis that are listed by these types of codes. The second types of codes that we talked about, which were procedure codes, Are also known and commonly used are HCPCS. So there’s two types of these codes. One are CPT codes, these are numeric and tend to cover provider services. The second type are all the services not covered by CPT codes and these are alphanumeric codes and include types of services like ambulances and any facility administered drugs. So chemotherapy drug that's not obtained through a pharmacy but it's administered by a physician might have a HCPC level II code that begins with a J. These are called J codes. There's many other types, but these cover any types of services that aren't procedures provided by a physician, which are coded by the CPT codes. Finally, there's NDC codes, which are used to identify drugs or pharmacy claims. These are ten digit numbers with three segments that identify the labeler, the product and the package. This could also be thought of as the manufacturer. The form and dose, say, a capsule in 100 milligrams. And the package that it comes in, in a 30 tablets or capsules. Related to all of these codes are diagnoses related groups which are bundles that are used for payment by Medicare and are adopted often by commercial insurers. For combining claims within inpatient, part A is the same in commercial claims as inpatient claims. And these are combinations of ICD-9 and CPT codes within a hospitalization, grouped together and then assigned a single payment. Note in this case, the position of the codes, so whether a code is first or second and when they occurred during the stay matter for the algorithm that are used to group the codes into the DRG payments. Finally, one point about the information in the claims data is that claims data does contain financial information, in particular, the charges and the payments for the services provided. because that's the point of the claims in the first place. But when using claims data to identify payments or prices, it's important to remember all of the elements that are in claims data. First, costs. Costs are not usually included in claims data, but the elements of claims data are often discussed in terms of costs. However, they're not in claims. Costs are the amount it costs to produce or provide a service. So if it costs $10 to buy lemons, $1 to buy water, and you make lemonade, the cost was $11. If you charge $5.00 per glass, that's the same as the charges and claims, those are the amounts billed by the provider. It's also sometimes referred to as the list price. However insurers, either Medicare or Medicaid which are payers or commercial insurers, either negotiate or set the prices they pay, which are not always, almost never equal to the charges. In this case, the payment, or the allowed amount, is the amount paid by the payer and the insured out of pocket. So, it may cost $1 to make a glass of lemonade, you charge $5, but you're paid 2.50. Now when an insured pays out of pocket, they can pay through multiple sources, sometimes these are broken down in the claims. The deductible is a fixed amount the patient has to pay before insurance kicks in. Coinsurance is a percentage that the patient needs to pay of the allowed amount, and co-pay is a fixed fee they need to pay. So, there may be a $5 copay for an office visit. An office visit payment may be $100, if the patient's deductible is $20 and the co-pay is 10%, then they pay the first $20. What's left is $80 then the patient pays 10% of the $80, and then that equals $8. So for that office visit in this case the patient would have paid 20 plus eight plus five. If they go a second time because they've met their annual deductible, they'd only pay the eight and the five. Well, sorry, they wouldn't pay the eight, they would pay ten which would be 10% of the 100 and the five. So the second visit would be $15. The first visit would be $33. Finally, expenditures are payments times the quantity which equals total spending. So the allowed amount times the total services would be total spending. So in our example it would be, of office visits it would have been 33 times one visit plus 15 times one visit, which equals total expenditures. So briefly, here's an example of what claims data may look like. So, here we can see that for a single patient. There maybe one claim. We know the age of the patient, how long they stayed in the hospital, and the diagnosis group that their claims were bundled in to. We also know something about their location, and we know the hospital payment. Which is then broken down into how much was received by the physician. As well as what the patient paid in coinsurance, co-payments, and deductibles. In claims data, it's also possible to see the same patient. So here where we I highlighted earlier, this is the same patient with two different dates of service. You can see though, it was in the same year so they have the same age. Previously we would also discuss diagnosis codes and procedure codes. So here you can see, That four diagnosis codes are listed but they're not always filled in. So for example, the first patient had all four diagnosis codes. But the second patient on their first visit only had two diagnosis codes recorded. But, both of those patients had five procedure codes listed. It's important to understand the elements of claims data when trying to use it to perform any type of technology assessment.
