The technology that has become useful in healthcare organizations is Computer-Assisted Coding (CAC). This technology has saved time and revenueforfforr the organizations and achieved productivity in the industry. Generally, with its introduction into the system, the coders thought of it as their replacement but soon with its implementation realized its effectiveness as an assistant. This paper is about the cost-effectiveness and the implementation of the CAC in the Healthcare Industry.
There will be a detailed outlook of what CAC is and its benefits to the organization. The benefits of the CAC implementation are many and the healthcare industries who have realized the benefits of such computer-assisted software have geared to progress in this fast-paced world. Now the days of tedious work are over for the coders who no longer have to waste time coding some of the simplest tasks. They have been categorized more as analysts for now the coders have more time on their hands to tackle the more difficult codes. This is all because of computer-assisted coding which provides results that are consistent and error-free.
The productivity of the healthcare departments increases with the use of the CAC as is seen in this research. The information that is gathered for the purpose of this research is qualitative as well as quantitative. The information is both primary and secondary in nature. The researcher has made extra efforts to negate any biased views in the collection and formulation of the information. The journals and books used in this research are from authenticated sources.
As defined by the American Health Information Management Association (AHIMA) Computer-assisted coding (CAC) is the application of computer software that results in the production of medical codes that are later reviewed. These codes are formed for the documents that are presented by the physician.CAC is also defined as computer software that has the capability of reading physicians’ documents and generates records automatically from the clinical documentation available.
Health organizations face problems of reimbursement and coder deficiencies. Moreover, their productivity was on the decline as well as they were facing revenue problems with the wrong billings. To counter this problem the health organizations have sought technology help and it is here that the application of CAC comes into their systems. Many realized the effectiveness of the system and as it was a programmed machine the errors generated by it were negligible to none. The coding could be achieved faster with increased efficiency. The technology improved compliance; accuracy saved time and proved to be cost-effective.
CAC is implemented in many organizations but in this paper, only the application of CAC in the field of healthcare will be discussed. There will be a detailed discussion about CAC and its benefits in regards to the general implementation as well as specific to the health organizations. It is seen that after its implementation the hospital’s benefit in the revenue department, as well as the process of reimbursements, is speeded up. The system was seen to save the time of the coders and this increased the productivity as the coders could concentrate and spend time on the more difficult tasks.
According to Jensen (1991), an international agency for research on cancer, the International association of cancer Registries the technology leaves the coder to finally select the final code from a list that the system develops for the coder review. Many errors of the coding are encountered when it is done manually and this is attributed to many reasons but with the CAC technology, no such errors occur as the machine is consistent with its work. The work guarantee is also increased with the implementation of the system.
The defined objective of a CAC is the division of the work between the coder and the computer. This is in accordance with the comparative advantage between the two. The computer being a machine is programmed to go through a number of files and a much faster pace than the human coder. It can match the answers in a more timely and accurate manner. Weil, & Regan (2007), argue that the coder just needs to identify the best possible code from the list of codes that the computer finally draws. In this manner, a CAC system saves a lot of time for the coder and also decreases the need for basic interpretations. As the final word lies with the coder hence it is very much essential for the coders to be highly trained in their domain.
The research is divided into different chapters. Chapter 1 consists of the introduction; chapter2 is about the Literature review in which the definitions and all the terms that are used are explained. There is a detailed explanation of the coding and its works before and after the use of the CAC system. Chapter 3 is about the methodology that is used in this Literature review. Chapter 4 consists of the findings and the conclusion. Chapter 5 consists of the summary of the research.
The aim of the research is to evaluate the total cost of the implementation of the CAC system.
Computer-assisted coding is used today for the effective management of the services provided by the web and also in sending the coded batches from one application to another. IBM currently obtains 75% of the shares being the middle man in the process. In the field of healthcare, the coders are responsible for the identification purposes of the different procedures and the diseases. Codes are also essential in the case of the process of billing. Coders work on multiple systems in the healthcare department. On one system they look for the codes that they research related to the field of medicine and on the other system they use for the saving of the documents that they encode.
Medical coding is a very responsible and crucial aspect of the healthcare departments as any inefficiency can directly incur $8 billion costs to the administration. Error in coding means that there will be an error in the billing which will lead to the wrong payments and this will eventually lapse the process of payments to the health professionals.
Another aspect of computer-assisted coding is that the market needs to be hand in hand with electronic record systems. This has resulted in the CAC market merging with the electronic records of patients. This means that the coding of the medical documents is very much dependent upon the availability of the documents to be present in the electronic format. The system needs to read the chart automatically. The system becomes smoother when the documentation is in electronic form.
According to Weil, & Regan (2007), the natural language process is involved in the linking of the code in the CAC systems. The codes that are generated are later used by Medicare and insurance companies. Natural language is also used to establish links between the different codes. However, there can exist automated generation of the codes as well. This can be achieved without the incorporation of natural language processing. This means that the physicians are under no obligation and may just dictate their diagnosis and the process of treatment in their said fashion. In this regard, they sometimes might be able to put in all the necessary details or might just omit some.
The codes are used to match the statement of the documents or physicians’ statements against that of the treatment. The best results are obtained when the system is fully utilized to match the best possible treatment against a given diagnosis with the help of the CAC software.
The importance of CAC
The importance of the CAC software is seen when the system is unable to search the freehand documents for the required text. CAC proves beneficial in the case where the system is integrated with electronic documents and can pick the wanted text quite quickly.
Another feature of the CAC which makes it equally important is the quality of data mining. Spriet, & Spriet (2005), argue that data mining in the field of healthcare is very important as the physicians can contrast and compare the different methods of the treatment with a specific disease. They can search the different documents available with changing the criteria and can in this manner enhance their research and the quality of their output.
The ability of the CAC
The CAC software has the ability to locate the billing and the diagnostic codes apart from the free text. In the opinion of Weil & Regan (2007), it can then use the codes and further structure them to be used later on by the billing department or for any research that is required at a later time.
Working of the CAC
The language used in the CAC is natural language processing (NLP). This is responsible for the alteration of the text in the form of data elements. CAC locates the meaning in the text that has medical implications, for example, pain or appendix. It further pinpoints the statements that have conditions like positive, small, soft, etc. Eventually, parsing occurs of the data elements in context to the system understanding.
With the advancement of technology and its applications in different organizations, many have realized the need for research in computer-assisted coding. The companies involved in it are the following:
US Bureau of the Census
The coding capability of the ACTR’s is being tested by the US Bureau of the Census against the benefits of its applications.
Australia Bureau of Statistics (ABS)
Many efforts have been utilized by the Bureau for the CAC system applications and utilization by the industries. A survey was conducted in the year 2000 in which the changes relating to the coding applications in the Labor Force Survey (LFS) were introduced. It was from this that the information was to be gathered through the people who responded, identifying occupations to an appropriate Australian standard Classification of Occupation (ASCO). CAC technology is also to be used for the coding of the industry and used as National Census. The CAC technology is considered to be more flawless than any other practice being used.
Statistics Canada is involved in the testing and analysis of all such programs that have relation to everyday tasks and are repetitive in nature. Statistics Data has utilized the CAC technology for the data files of the Statistics Canada Labor Force Survey that are a culmination from all the regional offices of statistics Canada. The codes generated are against the occupations and the industry. Furthermore, the codes are based on the Standard Occupation Classification System and Standard Industrial Classification (SIC) System. This has made this tedious job a lot faster and easier.
It was the year 1978 That the SIC system was first introduced and it involves two steps. One is that the screening of the information is founded on the location of part of words or the identification of some keywords. This is done so as to identify the information that might be lost in the second step. The first step is called the keyword system. The second step involves the matching of the keywords against the descriptions or the information that is available. This means that whenever there is a need for the formation of a code in the SIC then the information is matched against the information that is present in the library. If the match is complete then the code is formed.
Coding in the Industry and of the Occupation
The Malaysian Standard Classification of Occupation (MASCO 1998) is utilized for the Households Income Survey, Population and Housing Census, and Occupational coding scheme for Labor Force Survey. The Ministry of Human Resources published MASCO in 1998. For the Occupational Classification, there are four levels like unit groups, major groups, minority groups, and sub-major groups.
|Major Group||Clerical Workers||4|
|Sub-major group||Office Clerks||41|
|Minor Groups||Secretaries and Keyboard-operating clerks||411|
|Unit groups||Secretaries and Verbatim Reporters||4111|
|Secretary, word processing||41111|
Table 1: Use of codes in different occupations. (Spriet & Spriet 2005).
|Group||Fishing and its services||050|
|Class||Fishing, hatcheries, farms, and services related||0500|
|Item||Coastal and Ocean swimming||05001|
Table 2: Use of codes in different occupations (Spriet & Spriet 2005).
The open-ended questions require coding and it is from here that the need for coding in healthcare organizations arose.
Medical coding and Errors
The principal aspect behind coding in the field of healthcare is to avoid inaccuracies. For a detailed understanding of errors in the coding the following causes are explained:
Sometimes what is verbally said can be misunderstood or the description may not be clear. Even at times, it could be incomplete. When the document is written freehand there may be a problem reading it wrongly or it may not be written well and is hard to decipher. Physicians could also be writing their documents in their own language and so the very first thing to do in this scenario would be to have it translated.
Weil & Regan (2007), state that this again could result in a lot of misinterpretations. For example maybe what the physician wrote in his native tongue meant something else to what was translated. There could also be a problem with a physician who is trying to write in another language that is not fluent in. This could have a drastic effect on the documentation for the physician might be thinking something else but falters in expressing it when penning it down.
Jensen, International agency for research on cancer &International association of cancer Registries (1991), argues that making choices in coding is very common. But this is another cause of the error creation in coding. While the coder chooses the information that might be relevant could be lost. For this not to happen the coder should be extremely familiar with the coding dictionary as well as the terminologies that are used in medicine. Even with the CAC technology, the need for a competent coder is essential.
Steps of Coding
There are four main processes that are involved in coding. These are explained as follows:
- Text recording: The text has to be recorded as is said. This is a very good way of text preservation even for coding at a later time.
- Text Translation: This is the step where the text is translated. For the purpose of the translation, either the coder or the coding system must have knowledge of the words and also a complete understanding of the sentences. Sometimes what a human can understand with ease is a bit difficult for the machine to comprehend. It is the context that is highly crucial for the meaning of the sentence. This is a common observation with words that could have different implications depending upon the way they are used. Spriet & Spriet (2005), are of the view that this is especially true in the coding of the occupations where the industry is important and further his qualifications to ascertain his position. For example, just coding simple painter could apply to a person who paints the house and also to one who paints on canvas.
- Categorization: This is the phase in which the state is matched to the category. In this phase, it is often seen that even though the coders follow the same set of rules they often result in different results. For any issues, an experienced coder is used as the last word.
- Process Improvement learning: This is the last phase of the coding process and in this phase, the main thing is the establishment of links. Here the codes are further evaluated for the optimum running of the system. Moreover, this phase involves error segregation and the cause of the errors. Hence it is a tuning of the coding process in order to provide consistency of the results.
The clerical coding process
These are the coding experts that are responsible for the coding translations. They read, understand and translate the text with the rules of logic and finally come up with the resultant output of the response. The rules are kept in a reference manual to which the coders often refer. However, with time and expertise, these coders usually end up memorizing the survey test and no longer have the need to look at the referral manual.
It is to be noted that the manual referral and the code assignment often result in errors and is also an expensive procedure. For the data which at first needs to be extracted and then undergoes the examination for the determination of the result is expensive. The cost is high for the individual examination of each code, also are the costs involved in training and looking for qualified coders. Sometimes it becomes quite cumbersome to retrain the coders.
For these problems alone the survey industry has now sought to look at the options of automated coding as noted by Spriet & Spriet (2005). This is for multiple reasons like cost improvement, quality of the coding, and the consistency of the coding. When quality is to be controlled of a manual process the difficulties are many, and the consistency of the results becomes an even more difficult task to handle.
Automated coding and the survey process
The data is collected in two ways one is the open-ended questions and the other is the specific questions. The data is coded into predefined categories or else into the newly created categories. What computer-assisted coding offers is that the data can be coded as is received online. Computer-assisted coding is also used for the data that has been collected through self-administered surveys.
The implementation of computer-assisted coding is typically used for the support of the interviews where the data that is collected is simple to code as observed by Spriet & Spriet (2005). Often in these, there are very long lists for which the codes need to be generated. They are done against a one-to-one coding frame and are not very complicated to code. The benefit of computer-assisted coding is that the response when coded online can be reviewed by the respondent themselves. The data which is not present in the electronic form is then processed through computer-assisted coding programs to make it easier for the coder.
These systems make the work a lot easier by referring to all the databases for the identification of a code. In the opinion of Jensen, International agency for research on cancer & International association of cancer Registries (1991), when there are difficult coding tasks these systems become extremely useful to the coder as they quickly go through the multiple data entries to code a text response. It is the judgment of humans that is the most difficult of all things to automate and because of this reason, the final word is left with the coder to decide the best code amongst the machine-generated codes.
Based on the processing and the information rules by the system designer the computer-assisted coding software is able to generate codes without the assistance of the coders. There are however only a number of limited entries that the automated systems can code and sometimes the more difficult and problematic tasks are left for the coders themselves.
Automated Coding systems in use
Many coded systems are being used for market research as well as for the purposes of the surveys. Following are some of the major applications of these systems and will provide information on the working and implementation of these systems in the solving of the coding problems.
The U.S Bureau of the Census is the first that test computer-assisted coding software. The first attempt was made in the year 1967, which was for an economic census. This was done to improve the consistency and the quality of the industry response which was obtained from the small business setups. The system that was used was named the O’ Reagan algorithm. The algorithms that were tested were based upon the set of rules that were derived from the previous coded responses. Association maps of responses were made, these maps were responsible for producing the most suitable code.
From the very beginning of the tests that were carried out by the Census Bureau, it was proved that such automated systems had a lot of potentials. The system they used provided Industry results for coding similar to those that were achieved by the clerical staff. It was at the end of the year 1976 that the Bureau decided on having the industry and the occupation census done manually. They came to this conclusion because at the time they perceived that the costs that were involved in the implementation of such systems were much more than what was achieved. It was also somewhat true to the extent that all the responses that were gathered from the 1980 census had to be manually keyed before being coded.
Automated Industry and Occupation coding system (AIOCS)
The research staff of the Census Bureau thought of another way of devising an automated coding system and their decision was to implement a system that was based on the steps that a cleric used for the achievement of his/her goal instead of the word to code association map. The coding manual was used as the reference manual and the decision about the code was in accordance with the referral match. The system was successfully used in the 1990 census.
In this system, the contents from the referral manual that was used by the clerical staff were used and not the sample of the questionnaire response. The program used synonyms instead of words and the abbreviations were sought to form the coding manual. Any words that were spelled wrongly were changed and abbreviations were found from the coding manual.
PACE was another application that utilized memory-based reasoning (MBR) for the production of possible matches. The possible matches were then further sorted out by the system by comparing them against the scores that were set against each word to increase the efficiency and decrease the errors in the results. The system required extensive computations and a total of sixty-five thousand examples that were fed into the system then produced the 4.5 million characteristics. It proved to be a better solution to the AIOCS system as it could calculate 57% of the Occupational entries and 63% of the Industry codes. Compared to this AIOCS could only code 57% of the Industry codes to 37% of the Occupational codes.
Automated coding by text recognition (ACTR)
This was first worked upon in the 1980s. It is used for the different surveys. The system utilizes near match methodology. The matching database is provided by the survey staff which could be coded manual entries as well as verified survey responses. The system allows the coder to go through the databases within the working of an application. The programming logic is to be supplied by the end-user. The system up-gradation was made in the year 1996. This was a more portable system.
According to Spriet & Spriet (2005), by the middle of the 1980s, the French National Statistics and the Economic Studies Institute (INSEE) produced a system they named QUID. The updated version of QUID was then named SICORE. The database of this system consists of prior codes. The SICORE can be used with any platform as it is written in C language. Also, it is not specific to any industry and can be generally used for any text coding. The system works by breaking text into bigrams that are the two-letter combinations of texts. It uses logical rule processing.
This is the computer-assisted coding that is designed specifically for the CAI application and does online coding. The system matches the codes with those present in the reference text to come up with the best possible result. In case the word does not reside within the database then the interviewer is prompted to input a word that closely resembles the word.
Computer-assisted coding features and Issues
The expert system technology is used for the purpose of creating a machine that could mimic the working of a clerical coder. According to Spriet & Spriet (2005), these systems arete, red natural language translators. They have the capability of understanding the fed input. For coding purposes, the system depends upon the knowledge rules. For the reference code help of the expertise of the coders is sought. Initially, it requires the expenditure of a lot of time for the reviewing of each code and then modifying them according to the system requirements. The use of a reference database is a bit complicated as common phrases and terms are neither are nor used in the database. The parsing phase is in which the rules for the standardization are applied for matching purposes. Sekaran & Bougie (2009) are of the opinion that this parsing and standardization are very difficult and care should be taken when doing these procedures so as to have an error-free system.
The most difficult of all the challenges that an automated system comes across is the coding in context. Multiple inputs are required for the system to be successful in this regard. The knowledge is sought from the experts in the development of the expert systems.
Non Expert systems
These systems perform the coding in a very different way; the example of such systems is PACE.
Interactive online coding
This is a more technically feasible system. The option of best code in the system is achieved through the information that is provided by the responder. What the coders do in this system which decreases the overall quality of the result is that they feed in the information so that they can have access to the code quickly. Doing this they often bypass the whole information provided by the respondent and so the quality of the result is affected. For the recorded information it is always possible to review and look at the errors caused. With a recording effort, however, the capacity of the system itself is reduced.
Development of standards and the error measurement
The quality of the code is measured by the validity and reliability of the code itself. A code is said to be reliable if the same code is used by two of the coders for the same situation. The reliability of the code is also dependent upon certain factors. These are the quality training and the amount of training that a coder has undergone. The input data quality as well as the complexity of the code frame is also attributing factors of the coder reliability. The good thing about computer-assisted coding is that the information provided is more consistent and is not generalized so as to make the results more reliable.
Weil & Regan (2007), argue that systematic bias is another problem when there is a system that utilizes a certain behavioral pattern. This means that when the coder start relying blindly on the system and picks the code without giving it a second thought could be making the mistake of systematic bias. To counter this there is a method that is used by which the codes are matched against a reviewed list of codes. These codes are formulated by the expert coders who unanimously then decide a single code against each entry.
The performance measures for the computer-assisted coding systems are accuracy and production. The production rate is determined by the influx of all the examples to which the code is to be assigned. The percentage of the correct entries that are coded is termed the accuracy code. There are certain decision parameters in the automated systems that can exchange the product for the accuracy rate. For the calculation of the benefits of the automated systems, the calculation of the production rate is very important with hand in hand accuracy of the codes as is done by the clerical staff.
Sekaran & Bougie (2009), argue that the measurement of the performance is dependent upon application, environment, and the data, all of which can’t be controlled. The very crucial factors which affect the accuracy and the production are the quantity and quality of the response data, the source of data collection, and the code frame complexity.
Computer-assisted Survey system
It is not just any simple data that is to be entered. The computer-assisted survey system means that the survey is to be automated and its full integration is possible. The checking session of the answers is done in the interviewing section. With the culmination of the interview the correction is done and the interviewer then submits the error for data to the office through a modem. At the office, the complete data from the different interviewers of the survey are collected. This survey is then prone to analysis and in the case of the occupation survey, a little of the coding is also required. Sometimes in the more complex surveys, all of the corrections cannot be achieved in the interview session.
Coding of the Medical Information
These coders are specifically trained to understand the medical terminology. They code the documents in accordance with the guidelines that are set forth by the hospital administration that they are working for. The medical doctor is always present for any reviews on the coding or if there is any problem in the understanding of the cases. HE is made responsible for all the questions and reports to the pathologists and the physicians. Doing this is very crucial as it affects the efficiency of the system.
The input of the personal Information
The coders with the use of computer-assisted coding code the personal information.
Effects of Errors
In the field of healthcare, the effect of errors tends to place a report’s credibility in question. The coding that is filled with errors will lead to inconsistencies between the data tabulated, the physician’s document, and sometimes the conclusion itself. Inconsistencies in the tabulation though not a grave issue yet still such errors are irritating when reviewed by the supervisor. It could aggravate the situation if the errors are seen as negligence by the supervisor.
Error in the selection of a general symptom that is specific like coding tremor for Parkinsonism.
This type of error is most common in the coders who do not know a phrase to be coded. Such a coder usually fixes the problem by placing a code that is too general. For example, he would use alopecia instead of hypotrichosis.
Diagnosis, Syndrome, and Symptoms
Let’s say that a physician gave the Syndrome/diagnosis and also provided the symptoms. Now it is up to the coder whether he coded them both or not. By choosing to code them both the coder avoids the loss of information with the trouble of going over the document many times. For a progressive symptom, the coder may choose to use multiple codes or simply choose the most serious one.
Insufficient knowledge about the subject
Coders who are not familiar with the clinical terms might use a term that does not justify what the clinician had in mind. For example, the code might coin in hallucination, instead of the term mania.
Misunderstanding in writing
This is another common cause of error by the coder. Sometimes the handwriting is hard to understand as it is not clearly written with well-formed words. Or else a coder might read a word wrongly mistaking it for another similar-looking word. For example, phonophobia instead of photophobia could easily be confused.
There is always a risk whenever there is the need to translate before coding.
The errors in the coding can be detected by comparing the document against the coded one. Moreover, the codes could also be rechecked by the pharmacovigilance physician. If the coding errors in the report are in abundance then the work needs to be done all over again. Spot checks by the audit are also beneficial in targeting the work of a specific coder and reviewing it.
Ideal system Criteria for implementing Occupation and Industry Code
As a guideline on which to delve into the working and productivity of the CAC system, the Ideal systems are required as stated by Weil & Regan (2007). This helps in the provision of information that is required to input into the CAC prototype so that its fits the system it is to be used for. Mostly an ideal system is composed of certain characteristics like easy usage, must be accurate and could be relied upon, an identical match to be displayed first, could obtain information from an interview, could work at a fast pace, updating of the codes should be easy, have auxiliary fields for the restriction of search, be cost-effective, must have an allowance for interviews and have the integration of allowing automated or semi-automated coding.
Preventive measures for Quality Assurance
To tackle the problem of the errors and increase the quality of the result it is always a good practice to translate the document in the physician’s own country and then it can be sent for coding. Also is the fact that it is often stressed that the handwriting should be such that it could be read without errors.
Spriet & Spriet (2005), argue that the training of the coders should be synchronized in order to make them knowledgeable of handling similar problems and situations in a unified way. This inconsistency is evaded to some extent. Moreover, the coders must be responsible and qualified. They must have good knowledge of the use of medical terms and their meanings. The knowledge of the coding dictionaries is also a must. According to Weil & Regan (2007), a logbook is often a useful tool to refer to for guidance.
For the healthcare departments to manage the cash flow with speed and accuracy it is always beneficial to improve the accuracy and productivity through the coding processes. The process of coding is a complex and time-consuming process in which the coders need to be very careful. The CAC technology is now available to help the coders convert the medical information like billing, procedures, diagnosis, tests, and other services into coding for review by the revenue department.
Today for the implication of the CAC software, two options are available. These are the structured input and the natural language processing (NLP). Both are computer-assisted models and are based as such that at any time the result can be viewed and edited by the human coder. These are in contrast to the automated coding systems that with the intervention of the human coder send the code directly to the billing. However, the use of this technology is not in common practice for now.
Computer-assisted coding (CAC) is software that is used for reading documents that are available in electronic format. For the layman, it is software that is used much like the spell checker. The software can identify certain phrases and words inside the document with additional programming that allow the software to identify the parts and contexts in the document in which the specific words and phrases appear.
The working of the software is different from that of normal word-finding software. It has to be advanced in the sense to determine the codable condition from the noncodable condition. This is done through different computational methods. The software can be designed to interpret and understand the different language patterns and make decisions that are based on the examples seen in a previous document. What CAC software I able to do is that with the help of natural language processing ability it makes the computer understand the English language and decipher the reports that are given by the physician. It then nominates different codes for the different patients. How the software achieves the method of coding is that it learns to emulate the way in which humans code by statistical prediction and learning.
Initially, the CAC software could be used for chest x-rays, X- Rays for the knowledge of a fracture present or not, and routine mammography. But today the ability of the CAC software has exceeded by leaps. The use of the software is to a coder who no longer needs to be sitting coding thousands of mammography reports but utilizes the same time doing something more difficult in terms of coding. This is because the CAC software is quite capable of doing the mundane and the repetitive coding by itself.
Another aspect of this software is that they are gearing up the present coders to excel in quality control and auditing so that they can perform their tasks better. Today the coders work on more testing and difficult codlings as the routine is taken up by the software.
Natural Language Processing
Natural Language Processing (NLP) is an extension of Artificial intelligence which originated in the 1950s. It has the capability of electronically reading the dictated or transcribed files. The NLP software is also termed the NLP engine. For the identification, analysis, and derivation of the documents, the engines utilize computer-based reasoning or semantic rules. On the clinical side, the words are changed to codes like CPT and ICD. For every record, the codes are then placed in a queue for review by the coder. Time is saved for the coders who then just finalize the code from the suggested codes instead of doing it right from the very beginning.
In order to gain optimally from these systems, the documents that are used for the coding should have a standardized format. The document should be in a set order for the program to provide its optimum like for example the use of headers like History and Examination. NLP engines also provide their optimum when the system through which the coding is to be done can be accessed remotely possesses an interface. Spriet & Spriet (2005), state that the outpatient is the ideal environment for the application of these engines, because of the reason, that the documents which are fed into the system are template-based and the medical terminology that is used is not very vast. The engines have been seen to provide excellent results in the following departments:
The use of the CAC of this type is present typically in the physician’s office. It utilizes the use of clinical terms. The clinician records his document in a handheld, with the selection of words or phrases the clinician can obtain sentences formulation resulting in paragraphs. Procedures and the diagnosis are pointed out with one resultant suggested code.
Occupation coding techniques
The variants of the occupation coding techniques are:
Computer-Assisted Coding (CAC)
Automatic batch coding (AUC)
Manual coding is diminishing nowadays with speed. There are many reasons for this. It’s a world of speed where mistakes are costly hence these are some of the drawbacks which are not favorable when it comes to coding. Other drawbacks of this system are:
- System costs are very high
- Requirement of a number of coding staff which is difficult to obtain
- Inconsistency of the coder output
- The results are not always accurate and the quality is not high
The manual coders are burdened with many tasks at one time and as the burden of work is a lot they often tend to develop shortcuts. This is against their training yet still is practiced amongst the coders. The results of which are loss of information and the output also vary from coder to coder making the results inconsistent.
This is now the most sought-after technique as compared to manual coding. In healthcare organizations, it is used to assist the coders to enhance their efficiency and productivity. It is now being favored over manual coding and the reasons behind this are:
- Quality can be monitored in real-time
- The results are accurate and are of greater quality
- The results are more consistent
Weil & Regan (2007), are of the opinion that basically the working of the CAC is to get into the hassle of all the tedious work and create a menu of the suggested codes from where the coder can choose the final code. The results are always consistent with this system no matter the number of times a document is fed into it.
Automatic Batch Coding
The decision rules that are built into the automatic batch coding make the system decide upon the code and this is the differentiating aspect between CAC and AUC. Moreover, the AUC system has to be complemented either by the CAC system or manually (Statistical office of the European Communities, 1997).
Issues Regarding CAC
With fewer individuals present as coders now there is immense pressure to develop CAC so that it can handle even more coding than ever before. In healthcare, the process of Reimbursement is dependent upon the coding. The more efficient and correctly the coding is done the sooner the organization gets their money, any faults in the coding system can lead to delay in money deposits. Moreover, CAC can also be used for the identification of fraudulent activities.
MedQuist Inc did a study on CAC productivity in the year 2004. CAC technology by the name of CodeRunner was investigated by the organization’s coding staff. The documents that were used were from the Emergency department (ED) of the hospital.
There were a total of five coders that participated in the test. The records for the coding were taken from five different hospitals’ emergency departments. The coders were paired with the facility and were nominated 100 ED documents for coding. The results proficiency was based on the hospital’s internal compliance of accuracy and the coding and billing guidelines were also used as per of the hospital. The records that were used were electronically transcribed and were fed into the CodeRunner through a direct interface. After this, the NLP engine of the Code Runner processed the records of each coder and was subjected again after a couple of weeks for review by the same coder.
The productivity of the coder was measured for records that were coded for each batch. The measurement of productivity was further obtained in two ways. These were a number of codes per hour and a number of records per hour. Code per hour was determined by dividing the number of codes in each batch by the number of hours that were required to complete the batch of codes. For the measurement of records per hour, the total number that took the coder to complete the task was divided by 50.
The result that was achieved at the end of the study clearly showed that the productivity ratio of each coder was increased with the use of the NLP engine. From the average taken of the results of the five coders, it was evident that there was an increase of 96% per hour in productivity with the incorporation of the NLP engine. There was an 85% increase in the average per hour code ratio. The results were in accordance with the hospital’s guidelines.
|Records/Hr Without NLP||Records/ Hr With NLP||Improvement Percentage||Codes/Hr Without NLP||Codes/Hr With NLP||Improvement Percentage|
Table 3: Coding percentage with implemented CAC system (WWW.medquist.com).
CAC application and result
South Shore Hospital is another successful example of the implementation of the CAC technology. It’s a 282-bed teaching hospital IN Massachusetts. Its Emergency department gets an average of 77,000 visits. The hospital with all the expansions and workload had only two facility coders and so was suffering from high code costs due to all the backlogs. Many records remained uncoded for months. It was MedQuist’s remote CAC technology that came to the rescue. The hospital’s ED reports were electronically transcribed, and the coders did not have to spend time consulting any charts. Spriet & Spriet (2005), state that CodeRunner made it possible to present the coder with an available set of codes to choose from and saved time for the coder instead of working from the very beginning coder just needed to review and select one code.
The most important fact realized with this implementation of the CAC system into the South Shore Hospital was that before the use of this computer aid a coder could only code 12-15 records per hour and after the CAC implementation they could easily handle 18-21 records per hour increasing the productivity by 40%.
According to Spriet & Spriet (2005), the technology only reaps benefits and unlike the threat that the coders initially felt with the implementation of these systems, they now consider it as helpers which assist in their overall productivity. The coders’ in fact have been promoted from being simple coders to data quality analysts.
The Electronic Health Records (EHR) are picked up by the CAC systems and put through the process of coding. In this manner, it fills in the gap that was present between the records and the healthcare financial systems and transcriptions.
Apart from productivity, other benefits of using CAC are accuracy, compliance, transparency, and consistency.
When the coding result matches the guidelines of the organization then the CAC system is termed as being accurate. Further accuracy can be enhanced when there is the recovery of the misplaced charges that were not coded properly, differences in the audit, and decreased denials. The increase in the accuracy is beneficial to the hospitals as everything that is their due is recovered and identified and hence the hospital is able to collect what is rightfully their right.
The CAC system is a supportive and see-through coding process. It has no faults and can be checked at any time. The working of it is such that it must produce accurate results without any fault which helps the organization by saving on to time which otherwise in case of errors would have been consumed for rebilling and rechecking.
Weil & Regan (2007), state that the transparency of the system is such that it provides all the details for checks on the codes used. Any changes made to the demographics or the code itself can be traced and checked against. Furthermore, the system is made manageable with the provision of evidence for all that is done from the start to then end in obtaining the coding results.
According to Spriet & Spriet (2005), consistency of the system is the checks and proofs that are made of the coding resources against the guidelines that are provided by the hospital. This means that the results that are obtained from using the CAC must comply with the guidelines that the hospital has set. When the consistency level is high the results are more dependable and there is a gain of confidence in favor of the results. Consistency becomes all the more important where the coders are employed with different skill levels.
Features of the CAC
- Shortening revenue cycles and Enhancement of the Case Mix Index (CMI).
- Reduction in review and audit.
- Speedy review, clarification, and finding of scanned and electronic medical records.
- Provision of flawless integration with any kind of encoder.
- Reduction of the contract coding services and overtime of the coder.
- Accuracy and reliability organizations records.
- Increased productivity.
Installments of the system
Software as a service (SaaS) is used for the installation of the CAC. It can also be installed as client-server software. There is not much hassle with the local installments in terms of the SaaS installation. The most common modules of the CAC are production monitoring, coding review, auditing, management reporting, and coding automation. The interface is required to transfer the record into the CAC system. Typically XML and HL7 interface formats are used. The NLP is compatible with all kinds of document formats. In the presence of programming interfaces, the services of the Web can also be utilized. CAC output is obtained in coded records like ICD-9 and CPT. There is also the availability of other information like code linkage, patient demographics, payer demographics, units, and modifiers.
Consideration for CAC Implementation
Any such solution that flawlessly fits into the coding system is advisable. It should be compatible with all the administrative and clinical applications without any disruptions. It should be an efficient assistant in the more difficult tasks.CAC should provide consistency, accuracy, and productivity to the hospital. It should have the ability to check for any faults in the document with the elimination of the data that is irrelevant or repeated. It should be able to suggest a menu of codes from which the coder can choose. All the work should be handled by the CAC leaving the coder only to review and make the final decision.
The capabilities of the CAC technology should be made clear to the administrator as well as the coder of the hospital. What the system can achieve must be general knowledge at all levels so that they can fully grasp the system’s application as well as the benefits it provides to the hospital. Recognizing the importance of the CAC the coders must readily accept the changes that the technology brings to them. They must look at it as an assistant to help them achieve their job goals in an efficient manner. Sekaran & Bougie (2009) are of the opinion that the knowledge about the system aid for the administrator is important as he needs to feel confident about the results of an application that the hospital invests in its system. Only when he is satisfied and confident about it then he would report and talk about it to the seniors.
CAC technology can be used in the system in many ways. One is the NLP in which can be used with the physician’s records without any major changes. Then is the structured input whereby the system can be assimilated with the physician’s documents and the last is the concurrent coding which is a real-time process. This means that the changes to the document can be made anytime. But whatever the case is for the system to provide flawless accurate results the documents should be complete. Any information that is not available cannot be fixed by the system. The CAC system provides manageability and system auditing.
Prospective Payment System (PPS) has become very common nowadays. With this, the use of encoder software systems has become equally important. The encoder is basically software that enables a coder to type text. In response to the typed text, the encoder generates a set of codes which eventually helps the coders in choosing the one that best suits the situation. Although the increase in productivity and the reimbursement process has resulted in the increased sales of the encoding software. But for the complete evaluation of such systems, these results are not enough.
In view of Weil & Regan (2007), for the product credibility, the vendors have benchmark tests against which they compare their products. The benchmarks are used to evaluate the products for accuracy, consistency, and completeness which can also integrate into the system analysis and the management that is required for the product of the hospital. This product is the most essential for patient care quality. The software is also beneficial to a system in the sense that it can be used to further enhance and develop the coder’s skills.
Encoders are usually knowledge-based systems that duplicate the logic and structure of the knowledge. The encoder helps in reaching a decision that is made by the coder and does not make the final decision. The coding vendors now have affordable solutions for the ones that want to use it independently. This software proves beneficial for auditing when once installed. There is one essential thing to be mindful of while installing these software’s and that is the system up-gradation. This should be done at least quarterly to be up to date.
Alpha II Coding System
The Alpha II Coding System is the brainchild of the Meditech Company. This software is specifically designed for medical coding. It is the most widely used software in clinics, hospitals, physicians, and coders. The system is one single user is sold for about $1300. The multi-user License is sold at around $1,800. The add ons are also available for the software which can also be obtained online.
Other Medical Coding Software
Flash Code Expert
Unicorn Alpha II series
Mede Trac Systems L.L.C (Coding Software)
AMA’S Medical Coding Manager
Ingenix’s Encoder Pro Coding Series
The input of the information in the medical field
The information about the personal data is coded into the system. The information like name, marital status, occupation, sex, age, etc is coded. This information is recorded on the date of the disease being diagnosed. This method enables the application of the proof of the criteria like the patient being a Britisher. Names also provide usefulness in the manner of correlating them with notification forms. However, as the names can change due to marriage so the coders need to upgrade the system information often. In the end, after all the verification of the input data is complete a program check is run so that any data discrepancies are caught within time and amended.
Control of Quality
Quality control in medical coding has to be a consistent and continuous process. Visual inspection is done for any obvious errors. Visual check of the notification form compared to the computerized in for about topography or morphology is especially of grave importance. For the mounting of the microfilm copies inspection of both the disease and patient is done as a final check.
Computer check is made to enhance visual control. Logical errors are corrected in the coding. It is the responsibility of the expert coder who is also the chief of the registry to re-enter the logical information. (E.g. in the case of a sex change patient). The type of checks performed by the system are curative surgery be acceptable only in the case where the extent of disease is local or regional, Surgery is the basis of diagnosis in the instances of curative surgery, morphology and topography acceptable only after inspection, paired organ laterality specification even for the unknown, autopsy decision acceptable only in the case of patient expiration, histology be present when the diagnosis is of histology and cytology, date of diagnosis be equivalent to the date of death or current date, the occurrence of sex-specific diseases in the related sex.
After performing all the checks on the coding, one last check is made of warnings. Where there is suspicion of code repetition, a warning of duplication is made. Warnings are also issued in the cases of multiple sites of disease and for any rare combinations.
Updating and Filing
This is the up-gradation process of the main data in the registry. Routine printouts of the summary of the statistics are done. All the forms are to be made in microfilm with the allocation of 15 microfilms per person. Errors are mailed for renewal to the system registry. In case of zero errors, the microfilm is finalized and the original is destroyed.
A section chief who is a medical doctor is responsible for the registration. He /she is responsible for the smooth running of the system and for any decision involving the improvement of the system itself as well as the clinicians, pathologists, and computer systems.
For a maximum of 1000 notifications usually, a staff of five coders, one programmer, and four other clerical staff are present. All these staff members are specifically trained for their job allocation. Further, the coding and the classification are looked after by at least two medical professionals. The remainder of the staff acts as consultants.
Forecasts and Market opportunities
From the year 2007 till 2013 rapid development of the CAC technology is predicted. With the advance in technology, the number of procedures will also become much more than what is currently being used for the CAC technology. Until now the system is already being used by a number of hospitals and its trials have been completed a long time back. The systems today are able to take upon all the lengthy and tedious work that the coders used to perform. This caused a lot of delay and hassle. But not anymore for now the system has created efficiency and is a savior of time and energy.
Before the invention of this brilliant technology, the healthcare departments were solely dependent upon the coders and this resulted in 30% of their revenue for the collection and payment systems. Compared to this the banking sector had long started the automated billing system that took 3% of their revenue. Franzosi (1990), believes that it is soon to be realized that the healthcare organizations will catch up with the technology and a time will come when the more efficient will overcome the ones that have lagged in acknowledging the computer-assisted coding technology
The CAC technology is easy to incorporate into the healthcare system. The input can be obtained from a number of different sources like voice recognition output, notes that are types, and transcription services. It does not even require any templates. It has the capability to formulate appropriate codes through structuring the note. The system does not miss, falter or omit any of the notes and after checking through them the notes are further sent either to the auditors or the billing system for further review. Correct coding is very important, under coding can result in revenue loss, and over coding can have serious implications on institutional risk.
CAC is a typical example of hybrid technology as everything is measured on a scale. The natural language descriptions that are formulated by the clinician about the health of the patient are put against a scale to determine the extent of the illness. Hence the means are measurable as well as methodical. The clinician measures the illness as the patient measures pain against a scale. The scales that are used are physician-specific rather than being universal scales. This means that the scale is according to the physician’s own experience and in this way, his expertise is transferred from patient to patient.
The technology aims to incorporate the physician’s or clinician’s methods for describing a patient’s condition. These are all taken up in natural language and are formulated against a coding scale and the natural language coding software that the physician has control of.
According to Chute, Yang & Buntrock (1994), there is one prerequisite for the future development of the CAC technology and this is the quality production and evolvement of the NLP applications and NLP itself. Moreover,r all this requires an approach that is process-driven for quality assurance and software development. NLP is hence the key defining feature of the CAC technology.
Healthcare organizations will always be prone to compliance and financial risks unless they address the validation problems and put a stop to the claims being rejected. CAC is more efficient as it does not require the additional efforts of resubmissions of claims and time wastage because of reworking.
In the U.S alone the hospital’s number has been on the decline in the past 20 years. The future is of the IT and physicians and hospitals will be crippled without its integration into their systems. An example is Kaiser. They enjoy being ranked no 1 in USA’s healthcare delivery system just because they had the foresight of incorporating IBM into their systems.
The number of patients has increased as compared to before and this has resulted in more procedures being in demand. Today the healthcare delivery system is evolving and people want better and better healthcare systems. The markets worldwide are estimated to expand from 44 million in 2006 to 66 Million by 2013.
When implementing the CAC technology into the system it is very important to consider the cost-effectiveness of the technology. For this the two things important for consideration are the :
Coding load versus manual coding load
The cost of the CAC technology was compared to the other methods of coding costs in a test conducted in 1991. The test conducted involved the data input of forty characters that were for different variables. It took 45 days / 7.8 person-years for 40 data entry people to code all the data. Moreover, the coded data were further verified by a second coder. This took a total of 90 working days equivalent to 15.7 person-years. This of course was taken in accordance to 230 working days in a year.
For an estimate of the keying time, the length of the occupation strings was calculated. This was calculated to be 18 cm. As per this the keying time of the string for occupation was 7 person-years or 1620 person-days.
For CAC technology the time required for the coding of all the descriptions is founded on Precision Data. This is the software that is developed specifically for conducting the test. 400 codes per day were assigned by the coders. This overall rate that was tasked did fall less than average to what the coders could actually achieve because it is generally seen that as time passes the coder performance is enhanced. It becomes better day by day by practice and gain of confidence in using the system. This ranges from 24/hr in 1week to some 140 per hr in 36 weeks.
For these reasons the precise rate at which the coder’s work cannot be accurately judged but still by the observations obtained in the test, it was estimated at 500 per day. Based on this rate of coding it was further assumed that 30 coding stations could be made available and for a number of 1.4 million descriptions that were to be coded it would take 12.2 person-years or 2800 person-years. Calculating everything the total time was found to be 19.2 years 4420 person-days.
|AUC with CAC||10.7|
Table 4: Resource Requirements estimation (Spriet & Spriet, 2005).
The methodology that was used in the research of this paper was Qualitative as well as quantitative analysis. Different research methods were deeply considered of which Qualitative analysis was chosen so that depth could be attained about the CAC system and its implementation in the healthcare department. The issues that evolved during the research process were properly identified. The objective of any research according to Saunders (2003) is that the objective should be clear. For these purposes, necessary assumptions were made with the corrective adaptation of the most suitable methods.
The aim of the research is to find out the cost-effectiveness of the implementation of the system in the healthcare departments. The researcher through the chapter on methodology clearly evaluates the different methods used for the research providing critique where necessary. The types of data that are used for the dissertation will be elaborated on and discussed in detail. The primary data which is obtained through the interviews of the different people directly involved with project finance and corporate finance are recorded and then the observations are taken down in the print form. The use of extensive secondary data is also used which is from the material present in the books and journals.
The research answers were sought in an unbiased fashion. The articles and the journals that were used for the research were authenticated and from the authors that were unbiased. Moreover, the information was gathered from the researches that were carried by some of the hospitals in this regard. The results and the differences in the coding were analyzed between the coding that was done manually and the coding that was assisted with the CAC in the system.
Information was also gathered from the hospitals that had CAC implemented in their systems. This information was first-hand and was recorded. Also, the coders that used the technology were also questioned about the benefits of any of the technology.
The data was collected from two sources through the questionnaires that were prepared beforehand. One of the sources of the data was the organization that had implemented CAC into its system. The five major areas were covered in the questionnaire were: accuracy, consistency, Compliance, transparency, and productivity. The other source from which the concerned data was collected was the user of the technology that is the coder. The coder was asked questions about the system and its effect on the overall performance. All the questions were open-ended questions and were formulated in a manner so as not to lead the subject to the obvious answer.
The measurable deliverables were formulated from the data that was finally collected from the sources. The rate of success of these deliverables was solely dependent upon the questions that were put to the organizations.
A sample of different works of the coders was taken. A verifier that agreed or disagreed with the code was shown the different codes. If he disagreed then it was taken as a difference. These differences were of three types as follows:
Subjective: Without any outward violation the code differed on the opinion of experts.
Non Subjective: A direct violation of the procedures for coding and the code declared as wrong.
Procedural Change: a change in the coding convention rendering the code wrong.
Franzosi (1990), suggests that the difference in the rate is actually the measure of the disagreement of the code amongst the coders. Hence for the coding processes, a dependent verification check was done. A set of codes was verified individually by each coder and also the coding supervisor. The codes that were inconsistent were then marked with NS against that code. Some of the codes where the code was written like “Indian” became confusing of its relation to South American Indian or American Indian and hence were put as “S” against the code. With the coding procedural change, it took a lot of time for the change to reach across all the coders and even more time for its implementation. Codes that fell into this category were marked as “PC”.
The error rate was calculated by a number of coding errors dived against a number of coding actions.
Error Rate = Number of coding errors/Number of coding Actions
Three ways independent coding verification was chosen as the method for the monitoring of the effect of the productivity of the clerical as well as the computer-assisted coding. Samples were selected from the different cases. Three copies of the same case were made. Different code work units were established and the copies were distributed amongst the working units. When they had finished with the assessment the results were compared. These resulted in the three possible situations as follows:
- Three-way difference: all the codes that were checked were different
- Three-way agreement: The codes were the same.
- Minority/Majority situation: two codes out of three were the same with only one different or one same two different.
The researcher has tried to process and structure the dissertation so that it establishes a clear understanding of the topic under discussion. The purpose of structuring is to create inferences of the meanings of the different data that are put together for providing more meaningful information. Careful consideration was made while putting down all the information. The subject was much revised and reviewed with the consistency of moving ahead gradually with the dissertation.
According to Sekaran and Bougie (2009), research philosophy is where the researcher fully establishes the understanding of the topic. All the data and the information gathered provided enough material to formulate in-depth knowledge of the subject which facilitated the method of carrying out the research to completion.
Out of all the research philosophies the philosophy of positivism of chosen. This was because of the fact that the philosophy helps in the establishment of the real facts with the generalizations that are law-like. The data that is collected is for the development of the hypothesis of the research. In order to further dig into the idea and develop upon it, the hypothesis is further tested and the theory is developed bit by bit.
It was particularly kept in mind that the material found on the topic was not biased and the researcher took great care not to be influenced by anyone’s way of opinion. In fact, the research led to the pinpoint of certain facts that were researched, and the conclusions are drawn were solely on the data that was available without any personal preferences. The data that was collected was from reputable resources and the opinions of the authors were further checked for any bias to the topic. The articles were well researched and authenticated.
For the purpose of the research two approaches were available the deductive approach and the inductive approach. These two approaches are further explained below:
This approach is used mainly for hypothesis formation and development. When this is done the strategy is formulated to test the hypothesis. In the deductive approach, the personal opinion of the researcher about the topic is not to be considered. In fact, it has to be put aside and the literature is to be used as the guiding principles for the deductions and conclusions. In it, there is a quantitative measurement of the facts established. The data that is collected for this method needs to be extensive as generalizations need to be drawn from the information provided.
Here the theory is formulated as a result of the data analysis that the researcher makes with the collection of the data. The research of the inductive nature is opposite to that of the deductive approach and often the researchers following the inductive path are critical of the deductive approach. It involves the establishment of relations between the different variables as well as a bond with the thinking and opinions of human nature.
Crowther and Lancaster (2008), argue that Understanding the topic through information gathering is not the strength of the inductive approach. Also, it is not restricted to the formulations of any alternate explanations about the topic making the methodology confined to certain parameters. Unlike this inductive approach is much more flexible. The data gathering is often inclusive of a few samples and this is helpful in the context of the researcher being more concentrated and can achieve better focus.
In the opinion of Saunders (2003), another approach is to use both approaches as a combination. This is often termed as the back and forth movement because of the fact that sometimes the user uses one approach for the deductions and conclusions and then switches to the other approach for the other half of the research. This reverting between the two stages is very beneficial as it provides the researcher to optimally thrash out the complete understanding and in formulating the deductions through well-researched data.
The method of approach used in the research was both the deductive and the inductive approach. There was a need to gather a lot of information on the topic because opinions from the different businesses involved in the research were important as well as the focus was necessary so as to ascertain the understanding of the choice made by companies. There was a need for both a flexible manner of research as well as a rigid one in some contexts.
Both qualitative data and quantitative data were required for the research purposes. For the understanding of the different concepts, the use of qualitative data was very important, and to establish the facts the quantitative data assistance was used. Both the inductive and the deductive approaches were alternately used in the method of the research so as to gain enough information and then establish it as well.
Research Process and Data Collection
Qualitative data was used in which the well-researched topics of established authors were taken. The articles were unbiased and authenticated. Furthermore, there was also the use of the quantitative analysis where analytical tools were used for the establishment of the facts. This type of information for research purposes was gained from the different companies which were directly involved with project finance. This information was also used for the preparation of the interview and preparing the questionnaires as it provided insight into how the companies responded to certain problems and why they approached a method in a certain way.
The method of the primary research was through the interviews conducted of the different companies and the questionnaires that were made for information gathering purposes. The interviews could be face-to-face or over the phone, it could be structured or unstructured depending upon the relevance of the data that is to be used.
The interviews that were conducted were from the U.K’s well-known companies. For the authentication of the interview, the interview was recorded in the context of it being anonymous. To counter any biased nature there were no leading questions asked in the interview. As there was a time limitation all the interviews could not be conducted personally but instead were carried out over the telephone. The interviews were well prepared in advance for the proper guidance and conduction purposes. Secondary data was made available for the research from the reports and discussions of the different companies and the statistical analysis that were available on the topic.
Example of the successful Implication of CAC
Eastern Maine Medical Center is another example that has benefited from the use of CAC technology. According to the hospital, the processing speed of the records was increased to 15% initially which was a period of 45 days, and then another 15% within 90 days of its trial. The software used by the hospital was 3M CAC software. The software works by enabling the coder to choose a specific term (e.g nephritis) and then follow a logical decision tree.
This is done in complete confidence as the system guidance is very accurate. In the last part, the software allows the coder to choose a final decision for the final code. The program has the ability to enhance itself from the mistakes made. The system does this by adjusting itself according to the responses of the coders to its recommendation of a specific code. The system can very easily pick when the code it is flashing is not acceptable to the coder and checks itself. The adaptation of the system is useful in the training of it to the specifications of a particular hospital.
Involvement of the Staff
The hospital integrated the software successfully into the system by the active involvement of the coders in the implementation process. As the coders were able to see for themselves the features and their installment they became more confident of the system and of its features as well. The coders are very appreciative of the software’s abilities as it provides them relief from the most tiresome and mundane tasks.
The technical side of the implementation of the CAC software into the hospital required extra cost in setting up additional servers. These were required for the transportation of the transcribed reports as stated by Spriet & Spriet (2005). There was also a requirement of the Interfaces that were required to be written for the radiology, lab as well as billing system. There was also the establishment of the link between the software with the discharge transfer system.
It was observed by the hospital that the more electronic documents are fed into the system the more accurate the results were. The software fell short when it came to finding text in the scanned documents.
Radiology department output
According to Sekaran & Bougie (2009), inpatient facilities have lesser requirements for the CAC software than the outpatient departments. The Massachusetts General Hospital in the Boston area is another example that has benefited from the use of the CAC software. The technology has been integrated into the system since the year 2002. Initially, the system was able to code one single case in a period that ranged from 2-3 weeks. But with time the system is now able to code a case the very same day that it is presented. 30% of these cases are auto-coded by the system. This auto coding also fluctuates from a range of 30-60 % depending upon the use of the auto code by the departments.
The software is seen to be easier to use with maximum output where the codes are much the same. Tests like mimeographs and x rays are such examples where the CAC provides the greatest productivity rates. The hospital is happy because of the decrease in the denial rate. Coders are also happy in reviewing high-tech procedures like magnetic resonance and tomography and it provides them with more money-making opportunities.
The thing that is notable is that with the increase in productivity the hospital didn’t have to add more coders into their system. In fact, they were able to save their office space as the coders were shifted to work from their homes and this also resulted in saving the hospital money as office space is quite expensive to get in the Boston area. The hospital admits that the system takes a while in sinking fully into the system and patience is required but with time it reaps benefits that are really commendable.
From the data that was collected from the different researches that the hospitals carried out it was evident that productivity increased with the implementation of the CAC system. The time was saved for the coders who increased their usefulness in concentrating their efforts on harder tasks. The costs were lowered with the effective establishment of the system by speeding up the billing processes and saving valuable office space as the coders could work from home.
The data that needed to be collected for this research was not as much as the research could have made use of. There are as yet not many hospitals that have implemented the technology into their systems hence the observations were from the few that were available. This was the restriction and risk of this research paper.
The healthcare departments were facing a lot of challenges before the implementation of CAC into their systems. They had to depend upon the small influx of the coders who took a lot of time coding and sometimes because of the coding errors the work had to be done from scratch. It took a total of 30% of their revenue which was for the payments and the collection. Those who altered their systems in the wake of technology have survived major setbacks and Keiser hospital is one such example. CAC is the simplest and easiest technology that can be implemented into the system. The technology gains its input directly from the notes or the voice recognition system from the clinician.
Without the use of any further templates, the system can formulate codes. The good thing is that there are no human errors in it, it does not miss anything nor does the system falter in its working. The final word is still left at the coder’s discretion who can choose the best possible code from amongst the list that the system generates. As coding is very important to the health departments it is often seen that with the under coding they are affected through revenue lost and with over coding they face institutional risk.
Computer-assisted coding markets are predicted to jump from 50 million in the year 2007 to 3 billion in the year 2014. Of course because of the primary role of the insurance companies in the health department in the U.S the growth will be optimal there. The growth will be in conjunction with the awareness of the automated systems in the health departments as well as for the decrease in the overall costs that the hospitals bear.
Today technology has infiltrated every system and all organizations and businesses are handicapped without the use of technology in their systems. To meet up with the fast pace of the things in the environment each and every industry needs to implement changes within its structure to remain competitive as well as productive. It is for these reasons that CAC has solved the problem the coders used to face of logging everything manually and spending a lot of time coding the records.
The system itself has seen a lot of changes as the technology is being improvised to be more effective and productive in all the fields in healthcare. This rapid technology is rapidly evolving in a manner that is becoming more affordable as well as more accurate. With the help of the utilization of this technology will make any business more marketable.
Revolutionizing of the health care system is now just around the corner. The CAC software enables the practitioners to keep tabs on the workflow as well as compare the best available treatment/practice that could be applied. The software is extremely useful in providing the clinician or the physician with instant access to the patient’s treatment and diagnosis and provides an instant comparison of the different treatments for the same disease. The doctors can then follow up on the treatment provided and know of the 100% care provided to the patient. This process leads to a better health care delivery system. The department of radiology especially benefits from the system as they often detect conditions that later need a follow-up.
American hospital association. (1985). Hospitals, Vol 59, USA: American hospital association.
Crowther, D. & Lancaster, G. (2008). Research methods: a concise introduction to research in management and business consultancy, UK: Butterworth-Heinemann.
Couper, M. (1998). Computer-assisted survey information collection, Canada: Wiley.
Chute, C. G., Yang, Y. & Buntrock, J. (1994). An evaluation of computer-assisted clinical classification algorithms, AMIA annual symposium proceedings archive, 162-166.
Franzosi, R. (1990). Computer Assisted coding of Textual data, An Application to semantic grammars. Sociological Methods and research.19(2), 225-257.
Lee, R. M. (1995). Information technology for the social scientist, London: Routledge.
Porter, S. (2008). Tidy’s physiotherapy, USA: Elsevier Health Sciences.
Saunders. (2003). Research Methods for Business students, India: Pearson Education.
Sekaran, U. & Bougie, R. (2009). Research methods for business: A Skill Building Approach, USA: John Wiley and sons.
Spriet, A. & Spriet, T.D. (2005). Good practice of clinical drug trials, France: Librairie Luginbuhl.
Weil, D.A. & Regan, R. (2007). Independent Medical Coding: The comprehensive guidebook for Career success as a medical coder, USA: Rayve Productions.
Jensen, O. M., International agency for research on cancer, International association of Cancer Registries. (1991). principles and methods, France: IARC.