|Table 1. Common RIS flags.|
Hospital radiology departments began acquiring radiology information systems (RIS) in the early 1980s. Their purpose was to help monitor the efficiency of all phases of the process of imaging patients. At their core is an accession number automatically assigned when an examination order is placed. In spite of the maturity of RIS, radiology practices seldom take advantage of the accession number as a technology-driven tool to verify that they have captured and billed a dictated examination. This article will discuss how the numbering system could be used.
In order to explain how a radiology practice will use the accession number, it is appropriate to describe a few basic features of RIS. The first is the flag or hurdle designations that describe the sequential events in providing clinical service. Table 1 lists a series of common flags.
Bar-code technology helps expedite the assignment of the flag. At each hurdle point the responsible parties will wand the bar code on the document that accompanies the patient or film. This protocol is also made easier in departments that have a picture archiving and communications system (PACS). The flags that were linked to film movement and storage are replaced by other methods of confirming electronic retention of images.
Linked to each flag is a date/time marker. This is a critical component of the RIS, and its availability helps in the design of a monitoring system for examination capture, as will be explained later. The RIS continuously builds exception reports that pertain to the hurdles; the early ones are not as critical as those in the later stage of the process:
- Ordered but no patient arrival
- Patient arrival but examination not completed
- Examination completed but not dictated
- Examination dictated but not transcribed
- Examination transcribed but not electronically signed
The exception lists constantly change and are cumulative. Generally, they are organized with the oldest accession numbers at the beginning. The date/time markers are the foundations for the reports that help hospital and department management ascertain throughput efficiency and caseload intensity during times of the day and week. Time gaps between flags can be graphed in virtually any manner.
Tracking Examination Capture
How can the RIS features help a practice track its own examination capture? The answer lies in the construction of a RIS-based database that utilizes the appropriate accession number and date marker features. I will construct an example to illustrate the strategy. First, we will use an accession number sequence that has proven useful in this environment (not the only type used by these systems). It is a 10-digit number with the following structure:
- First two digits Year
- Next three digits Julian date of year
- Next five digits Date-specific sequence number
We start with all orders beginning at 12:00:01 AM, February 1, 2002. The universe will consist of 500 examinations; the system-assigned accession numbers will be the range: 0203200000 to 0203200499.
|Figure 2. A timeline table illustrating what happened to 500 hypothetical cases.|
On February 1, and thereafter, there will be examinations performed that were ordered on previous or subsequent days. Some, but not all, of the examinations ordered on February 1 will be performed on that date. Some will never be done because of cancellation. We will use this finite universe to illustrate methodology without getting lost in the real world complexity of a typical department. Figure 2 represents what happened to these 500 ordered cases.
February 1 is a Friday; some examinations were performed on that date but most were done during the following weekdays. Not all patients arrived in the department because of cancellations; some arrived but were not x-rayed because the technologists cancelled examinations due to medical or logistical reasons. All completed examinations were dictated and transcribed by February 5, but a few were not electronically signed as of 12:00 PM.
Before describing a database extraction strategy, it is important to address compliance issues. From the insurers’ perspective, in dealing with claims from a radiologist, what is the date of service? Is it the date the radiograph was completed, the date the radiologist dictated the examination, or the date of electronic signature? As you see from the table, it could be three separate dates. Fortunately, the RIS time/date feature retains all three. Because insurers want to be able to audit the files of both providers (hospital and radiologist) to verify that a claim for a procedure is legitimate, they want to see the same date on both claim submissions. This suggests that the date the examination was completed should be the “date of service” on the radiologist claim, even if the dictation/electronic signature date is different.
The next important issue is the source of the claim information for the radiology practice. It is critical that CPT-4 code assignment come from a copy of the report, and not from the procedure codes of the RIS. Some billing systems seek to take advantage of department-assigned procedure codes by securing this data electronically from the RIS. This is a big mistake.
First, a radiologist cannot bill for a procedure unless it has been electronically signed. Some hospitals do not put in safeguards that prevent transfer of the RIS file until the electronic signature flag attaches to the accession number. This implies that a radiologist could bill an examination that was performed but not dictated/electronically signed.
Second, most RIS use alphanumeric procedure code designations because they are easier to remember than a CPT code (Ch2 versus 71020). Hospitals are less concerned about coding accuracy because they are often compensated differently. This suggests both a potential translation issue in the conversion of the RIS codes to CPT by the billing system, and possible differences between the actual examination performed and the RIS code designation.
The use of the reading is critical to the audit and compliance process because:
- The document is the true foundation for the clinical service of the radiologist.
- It contains the accession number as part of the header.
- It lists all the dates pertaining to the flags: examination completion, dictation, transcription, and electronic signature.
The process of extraction of critical field information ranges from complete manual entry to electronic scanning of the source documents. The evolution of scanning technology is moving the process away from labor-intensive capture of clinical and demographic account information. It will probably not completely eliminate the need for human intervention, but will handle most routine caseload.
The reason this is cited here is the additional work to capture the accession number(s) from the reading document, and making sure they link to the procedure(s). Billing companies do not like to add a manual step to the data acquisition process. There has to be a good business reason. The use of the accession numbers would reduce the need for staff checking readings against a department log (assuming they are even performing this step). This manual process would be replaced by a technology-driven alternative.
Accessing the Files
|Figure 3. C flag database.|
We go back to the timeline table to sketch out the RIS database we need in this process. As an auditor, my preference is a database of examinations that were assigned the “C” flag for each date of service:
Yes, this is not the date of the dictation or electronic signature but there are two important reasons for choosing it:
- It is the date of service that will be used by the billing system. This means the billing system can produce an electronic file of all accession numbers by the parallel date of service.
- It represents the entire universe of examinations that should have a dictation, regardless of whether it happens. On the radiologist’s behalf, it is important to know this.
The hospital must be willing to produce an electronic file of accession numbers based on the date that the “C” flag was linked to them. A practice with a good working relationship with its hospital might convince them to devote the programming resources necessary to build this daily file free of charge. At worst, the programming cost would be borne by the practice, and it should not be extensive.
The implication is that the RIS and billing system can build identical files of accession numbers by the same date of service. The billing system file will take longer for a few reasons:
- A copy of the reading will not be delivered to the billing system for many days after the examination was completed.
- The billing company will not activate the charge until it secures the demographic information, and verifies its accuracy. This could take as much as 2 weeks after the date of service.
It would be counterproductive to attempt a database match too early; it should not begin before at least 3 weeks from the date of service. Basically, the RIS and billing system files have identical fields: date of service and accession numbers. Database programs already have standardized queries that identify records in one file/table that are missing in another. In this case the RIS file is the complete universe of cases for each date of service. The goal is to find accession numbers in the RIS file that are missing in the billing system file. Ideally, there will be a perfect match. The worst case scenario should be a small number of missing records. The accession numbers tell the billing company the identity of the patient, and they access the RIS to secure a copy of the reading or discover that there is no reading for some reason. This process is valuable to the radiologist as a safety net concerning breakdowns in the department or billing system information capture.
There are impediments pertaining to interventional procedures and radiation therapy. My suggestion for interventional cases is that billing systems establish a modifier field for alpha characters to account for the larger number of codes that link to the original RIS accession number. The RIS might have a single number for an interventional procedure. That case contains three to eight surgical/diagnostic codes to represent the work of the radiologist. The modifier field would contain codes “A,” “B,” etc to represent the additional transactions. It is not certain if this technique will work for radiation therapy accounts because the method of coding physician services varies so much from hospital transactions.
Billing hospital radiology procedures is more complex than most other specialties (pathology and anesthesiology, also hospital-based, share similar issues). Radiologists have to rely on the hospital for all critical information necessary to file claims to insurers. This specialty submits more claims to insurance companies than any other. The cash value per procedure is relatively low. Billing companies have to keep their processing fees down to secure and maintain client bases. Otherwise, it would not make sense for radiologists to bill a separate component, especially in an environment of shrinking payor fees.
Billing companies seek to acquire information in the most cost-effective manner, but recognize that it may not be possible to capture every billable transaction. Many of the methods used to verify transaction capture have been in use since the 70s. The rationale is that it is too expensive to put controls in place that reduce the loss factor to zero. The incremental cost cannot be covered by the processing fees on the cash value of the small number of cases otherwise lost.
While radiologists generally understand this reality, it does not ease their anxiety. The billing companies have done a poor job convincing radiologists that the majority of billable examinations have been captured. The use of this technology-driven technique should be a more cost-effective way to verify examination capture. It is not a layering of a new process at a higher cost to the billing system. It is a redirection of effort that takes advantage of the department’s inventory control system.
James A. Kieffer, MBA, is president of a management consulting business with an emphasis on radiology practice management and strategic planning. He can be reached at (978) 287-0003