Chief Operating Officer/Principal, Artificial Medical Intelligence, Inc., Eatontown, New Jersey
The United States healthcare system is challenged with increasing regulatory pressures, accelerating bio-medical and clinical technological advances, tighter more restrictive reimbursement from payers and the payment exposure from an emerging patient liability financial class which is rapidly growing. This leads to the shrinking of funds available to expand a hospital’s infrastructure. It often results in healthcare information technology priorities and attention being limited to the primary financial application platform and the clinical information system (CIS). Millions of dollars are committed to revising an old legacy CIS, often decades beyond its end of lifecycle. The assumption is that a new homogeneous CIS environment with electronic record capability will be the enviable byproduct of any upgrade initiative. Unfortunately, this is seldom the case as these systems have not yet been able to encompass the entire healthcare clinical-revenue cycle systems.
Since most clinical information systems parallel order entry systems in the business vertical, health information management (HIM) departments consequently have been put in the position as a secondary priority within hospital systems. The CIS systems that are available to hospitals have yet to include a fully functional Computer Assisted Coding (CAC) capability within the main system. Therefore, CAC technologies have been historically made available from coding specialized companies that offer bolt-on systems that complement the CIS system. It is interesting that most CIS systems have not made it a higher priority to provide an integrated coding system tucked into their platforms. The process of efficiently and accurately coding is a vital part of the revenue cycle and financial system because of its interrelationship with the healthcare billing process, which is the epicenter for hospital profitability.
Currently, despite the primacy of medical coding to the hospital revenue cycle and with CAC maturing and refining according to Moore’s Law mathematical rates,1 these systems are certainly not viewed as a priority for new systems purchase.2,3 The concept that is accurately coding a patient’s chart affects the hospital bottom line is not a new concept.3,4,5,6,7 However, due to the relative newness of CAC, real return on investment studies that can be generalized, justified and expanded are rare and typically incomplete.8,9
Despite the crucial role medical coding plays in the hospital financial universe, and the inefficient, burdensome manual process that coding presently involves, it seems that CAC solutions of any type nonetheless remain low on the radar screen. There is indeed a compelling argument for the immediate deployment of CAC systems throughout the industry based on the very nature of the present day manual coding process.
The level of acceptance of CAC technology is further hindered by the muddied waters of complexity fomented by a few early CAC developers who have existed in this field for a number of years. Their methodological approach has been widely trumpeted and reviewed by the existing HIM community as having limitations, due in part to the fact that many of these solutions can only code within particular specialties such as Radiology and or the Emergency Department.8,10 Alternatively, the need to improve clinical documentation often by leveraging a template driven system, often viewed by the clinical user as confining and limiting, also slows CAC acceptance. Unfortunately, those technologies that have technological limitations further defer widespread acceptance of the technology until these constraints are alleviated.11,12,13 One only has to recall the Apple Newton’s early failure contrasted to the overwhelming success of Palm and other PDA’s as evidence of this phenomenon.
The Newton was a series of personal digital assistants developed and marketed by Apple Inc. An early device in the PDA category—the Newton originated the term "personal digital assistant"—it was the first to feature handwriting recognition. In 1985, after Apple discontinued its poorly selling Lisa computer and faced plummeting Macintosh sales, their CEO, Steve Jobs, was ousted from the company he started from his garage. By 1997 though, Jobs was back at Apple, and picking up the mess. The company was months away from bankruptcy. Jobs turned the company’s focus to innovating great products again. By 2012, after inventing the iPod, the iPhone and the iPad, Apple became one of the most valuable company’s on the planet.
Furthermore, there is a continuing Achilles heel within the medical coding community involving coder accuracy and variability.14,15,16,17 Even before the arrival of CAC systems, there has been no "gold standard" within the coding community.17 However, the appearance of CAC solutions brought into sharp focus the accuracy question. CAC solutions were forced to wear the "inaccuracy" label from coders, when, in reality, coders themselves admit that coding often has no real universal "right answer."17,18,19 Typically, for any new breakthrough to be accepted there is always the additional challenge of creating a new paradigm. However, with CAC, there seems to be an unusual number of obstacles.
CAC Negative Issues and Obstacles
- Not a hospital technology priority
- Minimal data on return on investment
- Perceived complexity of implementation
- Technological limitations in a clinical setting
- Specialty and subspecialty- specific
- Accuracy in a world of coder variability
- Staff replacement
CAC systems have the potential to decrease the inefficiency and variability pervasive in the manual coding process. To date, there are numerous CAC deployments even with real limitations in some of the older solutions available. Some of this lack of acceptance appears to stem from the self-stated complexity in the methodology creating a negative mindset hindering the implementation of these systems.15 New data suggests that CACs can easily serve to improve HIM process flow and document handling efficiency. Today the new CAC systems can process any document from any source in any format, irrespective of admission status, location or type of service. These new technologies show an even greater benefit with other more involved and complex types of coding transactions, such as found with ICD-10. Even though new code sets typically take coders longer to process, with CACs it significantly reduces processing time and time reduction saves money. This provides an easy financial justification for this new type of CAC solution.
Another volatile issue is the coder’s concern about job security. This, no doubt, also adds to the resistance to CAC deployment. This is a false misconception as coders leverage CAC’s to quicken the coding process and therefore focus their time and energy on auditing charts to make sure they are accurate. The roles for coders are becoming more pervasive and complex as new requirements and regulations are enacted. This will be the evolution within HIM, directly paralleling the history of technological innovation within our society. For example, today instead of using typewriters to create correspondence, we use computers. All indications point to a greater acceptance of CAC technology in the future. The advantages of this evolving technology are listed below.
- Improve coder efficiency
- Decrease coder variability
- Streamline document handling process
- Connecting through centralized interfaces
- Maintain clinical environment neutral
- Coder redeployment to more mission critical tasks
Computer Assisted Coding, along with the evolving technology of Natural Language Processing, continues to become a more common HIM technology solution.10,20 CAC solutions now can provide a variety of innovative and useful abstractions of clinical information which can be seamlessly provided to healthcare workers, clinicians and patients alike, thus improving the delivery and efficiency of healthcare. CAC solutions that can provide analysis of the full array of medical documentation can play a focal role in managing the ever increasingly complex requirements facing the hospital HIM department. The promise that this underlying technology can be viewed as the next leap in the healthcare information age is promising and healthcare executives should stay tuned to the upcoming developments.
1Moore, Gordon E, Cramming more components onto integrated circuits, Electronics, 38:8 (1965)
2Survey: Computer-assisted coding not on the horizon-yet, Medical Record Briefing, (2005)20:5 http://www.hcpro.com/content/47018.cfm (4 of 4)5/3/2005 3:54:54 PM
3Featherly, Kevin. “News & Trends, ICD9 CM An Uphill Struggle.” Healthcare Informatics Oct. 2004: 1.
4Hsia, DC, WM Krushat, AB Fagan, JA Tebbutt and RP Kusserow. Accuracy of diagnostic coding for Medicare patients under the prospective-payment system, N Engl J Med, 318:6 (1988) 352-355
5Rode, Dan, Coding accuracy ultimately affects the bottom line - coding on healthcare providers’ uniform billing forms, Healthcare Financial Management, Nov 1991
6Fisher, Elliott S, Frednck S. Whaley, W. Mark Kmushat, David J. Malenka, Craig Fleming, John A. Baron, David C. Hsia, The accuracy of Medicare’s hospital claims data: Progress has been made, but problems remain; Am J Public Health (1992) 82:2; 243-248.
7Gall, Daniel W, Coding Accuracy of the Ambulatory Data System: A Study of Coding Accuracy within the General Internal Medicine Clinic, Walter Reed Army Medical Center, http://handle.dtic.mil/100.2/ADA372083, April 1998
8AHIMA e-HIMTM Work Group on Computer-Assisted Coding. “Delving into Computer-assisted Coding” (AHIMA Practice Brief). Journal of AHIMA 75, no.10 (Nov-Dec 2004): 48A-H
9The Impact of Computer-Assisted Coding: An Industry Whitepaper. MedQuist Inc., 2005. http://www.medquist.com/coding
10Testimony of the American Health Information Management Association to the Standards and Security Subcommittee of the National Committee on Vital and Health Statistics July 26, 2005 http://www.ncvhs.hhs.gov/050726p6.htm
11Friedman, Carol and Stephen B Johnson. “Natural Language and Text Processing In Biomedicine.” Biomedical Informatics. Ed. Edward H. Shortliffe and James J Cimino. 3rd ed. New York: Springer, 2006. 333. Rpt. in Computer Applications in Health Care and Biomedicine.
12Audet, Anne-Marie, Michelle Doty, Jordon Peugh, Jamil Shamasdin, Kinga Zapert, Stephen Schoenbaum, “Information Technologies: When Will They Make It Into Physicians’ Black Bags?” Medscape General Medicine, Dec. 7, 2004
13Bowman J, Stanfill, M, “Physicians cast a wary eye at Computer Assisted Coding” Journal of the American Health Information Management Association,75:8 (2004): 76-77.
14Morris, William C., Heinze, Daniel T., Warner, Homer R., Primack, Aron, Morsch, Amy E. W, Sheffer, Ronald E. (2000) Assessing the accuracy of an automated coding system in emergency medicine. Proceedings of the 2000 AMIA Annual Symposium, 595–99.
15Morsch, Mark, Heinze, Daniel, Byrd, David (2004) Factors in deploying automated tools for clinical abstraction and coding. Paper presented at IT in Health Care: Sociotechnical Approaches. Second International Conference, Portland, Oregon.
16Resnik, Philip, Niv, Michael, Nossal, Michael, Schnitzer, Gregory, Stoner, Jean, Kapit Andrew and Toren, Richard, Using intrinsic and extrinsic metrics to evaluate accuracy and facilitation in computer-assisted coding. Proceedings of the Computer Assisted Coding Conference. (2006) American Health Information Management Association.
17Lorence, Daniel P. and Awad Ibrahim. “Disparity in Coding Concordance: Do Physicians and Coders Agree?” Journal of Health Care Finance 29, no. 4 (2003): 43
18Friedman, Beth. “Waves of Change: Computer Assisted Coding the Latest Tech Tool for Coders.” For The Record 18.3 (Feb. 2006): 29.
19Kukafka, Rita, Bales Michael E, Burkhardt Ann, Friedman, Carol, Human and Automated Coding of Rehabilitation Discharge Summaries According to the International Classification of Functioning, Disability, and Health ; Journal of the American Medical Informatics Association Volume 13, Issue 5, September-October 2006, Pages 508-515
20Hripcsak, George, Carol Friedman, Philip O. Alderson, William DuMouchel, Stephen B. Johnson and Paul D. Clayton. “Unlocking Clinical Data from Narrative Reports: A Study of Natural Language Processing.” Annals of Internal Medicine 122, no. 9 (1995): 681–88
Stuart Covit1, M. Elliott Familant, PhD1, Cecilia Hilerio, RHIA2, Joshua Bershad, MD2,3 , Andrew B. Covit, MD1,2,3, Artificial Medical Intelligence, Inc, Eatontown, New Jersey1, Robert Wood Johnson University Hospital2, UMDNJ-Robert Wood Johnson Medical School3, New Brunswick, New Jersey.