This is an example of an excellent take-home final that I received from a student. The citations for the articles that I gave them to read, and subsequently asked questions are here.

Question 1

The table "Providing Dialysis Treatment" should contain the following attributes: Primary Key - Treatment Number, Foreign Key - Patient Number, Foreign Key - Nurse Number, Nonkey - Treatment Date

Based on the article, I believe that at a minimum the following information must be retrievable from the proposed AIS both in total and broken down by treatment type: The number of patients, the number of treatments, the revenue, the standard and episodic supply costs, the facility costs, the administrative and support staff salaries, utilities, the communications systems and medical record-keeping costs, RN salaries, LPN salaries, nursing administrative and support staff salaries, technician salaries and durable equipment costs.

I have assumed that supplies are identified specifically as being either standard or episodic. Therefore, these attributes are a part of the supplies table. Determination of supply costs involved in HD treatment and PD treatment requires relating treatments to supplies via the concatenated "Supplies-Treatment" table. This table has the treatment number and supply number concatenated into a primary key. In addition, the quantity of each supply used for each treatment can be found in this table. The number of patients broken down by each treatment type and in total is easily determined through an SQL query that relates the table "Providing Dialysis Treatment" to the table "Patients" via the Patient Number foreign key. The number of patients would be derived by a count of unique patients that were treated in the period being questioned. The segregation between HD and PD treatments is obtained through the "Patients" table, which would have a field specifying the treatment type that the particular patient is using. The number of treatments can be found similarly by a count of the Treatment Number attribute for the dates in question. To break the treatments down by treatment type requires linking to the "Patients" table. The revenue may be calculated by dividing the HD treatments by three and multiplying by the HD revenue rate and by dividing the PD treatments by seven and multiplying by the PD revenue rate. Similar to the determination of supply costs derived via a concatenated table, the nursing resources expended for each treatment are derived through the concatenated "Treatment-Nurses" table. This goes beyond what is required by the article, since this design will enable the manager to break down the nurses contributions to each treatment type by the hour rather than per nurse. I would encourage the clinic manager to further refine the cost allocation between the RNs and LPNs. My concern is that if the current activity driver for HD is increased by 0.5 for RNs or 1 for LPNs, it would cause the HD treatment to be unprofitable. An hourly breakdown would decrease this sensitivity and would provide trend information for average treatment delivery times for specific patients or nurses. Other information required is not specifically related to the table "Providing Dialysis Treatment". For instance the facility and utility costs are determined in a separate portion of the system. Salary information would be provided through the payroll section of the system. In some instances where the activity drivers are based on the number of treatments or the number of patients, the queries explained above would be required to calculate the applicable amount that applies toward each of the two treatments. In the case of dialysis machine operations, the assumption that the HD treatment requires 100% of these costs, makes the allocation simple by assigning all technician wages to HD costs. Beyond the minimum required information as mentioned above, the manager will certainly want other information such as the balance of accounts receivable in total or by provider. These numbers are easily found through the relationship between treatments and patients, the relationship between patients and providers and the relationship between providers and cash receipts.

For system control it is essential that each treatment be valid and receive a unique number. This number should be automatically assigned in sequence by the system prior to administration of treatment. This ensures that the number of treatments performed will be reliable. This is critical since revenues depend on the number of treatments provided and the treatments percentage is an activity driver for several cost allocations Additionally, each treatment must involve one and only one patient. To be a patient the individual must first be approved. The approval process involves acknowledgment from the provider that the prospective patient is covered by their plan. Therefore, if an individual is not covered they will not be allowed as a patient. This is important, since treating an uncovered patient is like taking on bad debt. The treatment date is crucial, since whether or not the treatment is booked as revenue in the period is based on the treatment date falling within the period. The nurse number is important under my proposed plan to directly allocate nursing costs to either HD or PD treatments based on the hours that each nurse spends on treatment giving. This will also more accurately reflect whether some nurses take longer in administering care or possibly whether some patients require extra time for treatment. As a secondary check for the payroll department, each nurses' hours per day can be derived under this system.

Question 2

Hospital and clinic managers face a number of threats in their decision-making process. Some examples of threats that exist are: Discontinuing a service that is profitable due to arbitrary cost allocation, continuing a service that is truly unprofitable, entering into a long-term capitation contract with a bid that is low enough to get the contract and also too low to make a profit or missing out on a capitation contract because the costs involved were not known well enough to bid at a reasonably low level that still provided ample profit. The exposure related to each of these threats is the potential lost profit and the potential bad will generated if specific treatments are discontinued and patients are inconvenienced. The risk of any threat occurring is inversely related to the quality of the information that is available for making the decision. Intuitively we know that less reliable information entails a higher degree of risk and uncertainty when applied to the decision-making process. To summarize the situation as it applies to the article, the threat is that of making a bad decision. The exposure is the total amount that could be lost on a bad decision including intangible costs such as bad will. The risk decreases as we create information with better underlying assumptions. To determine whether to pursue information with better underlying assumptions, we must first determine how much it will cost to improve the information to the level desired. Then we compare the net decrease in the expected error costs to the information improvement costs. If it costs less to improve than we expect to save, then improving the information is the right decision.

The clinic described in the article used a three tier approach, which took them from the simplest cost allocation model with a minimum of analyzed costs to the most sophisticated cost allocation model with all of their costs analyzed except for those of durable equipment. Based on the article we would conclude the following: Under RCC we should keep both treatments and HD is more profitable than PD, under M-ABC we should consider no longer offering the HD treatment because it is not profitable and under the H-ABC we should keep both treatments and PD is more profitable than HD. The cost benefit analysis is difficult, since in actuality we don't know what decisions would be errors and which decisions would be appropriate. Secondly, we need a basis from which to determine our exposure for making an erroneous decision. For the sake of performing a cost benefit analysis, I will make some assumptions that could be more erroneous than those on which the three methods are based. The assumption is that the H-ABC allocation method yields the results that would have been found had all the costs been directly traced to each treatment type. In other words the H-ABC method numbers are correct and the RCC and M-ABC costing methods are incorrect. The exposure under H-ABC is essentially zero due to the assumption. Therefore, the first cost benefit test is to determine whether the costs of providing H-ABC information exceed the expected costs of the RCC method. The exposure of the RCC method based on making bad decisions, would be to phase out the number of patients on PD treatments since it is deemed less profitable. If the clinic could convert all patients to HD treatments, they would anticipate an increased profit of approximately $41,000 (20,624 * (3/7) * ($9.17-$4.53)). Based on H-ABC, they would actually show decreased profits of approximately $78,000 (20,624 * (3/7) * (10.06-1.23)). Therefore, their exposure is $78,000. If we assume that the risk of making this decision to convert patients to HD treatments is approximately 20%, the expected cost of the RCC method is $15,600. If the cost of implementing H-ABC is less than $15,600, it would be appropriate to do so without even considering the intangible costs of the RCC method. The second cost benefit test would be to determine if the additional costs of converting from M-ABC to H-ABC exceed the costs of making the wrong decisions under M-ABC. The information provided by the M-ABC method would most likely lead us to discontinue the HD treatments. Therefore if we were to make this decision, we would be losing $17,627 in profits. This is the exposure, not including the intangible consequences, associated with the erroneous decision based on the M-ABC method. If the risk were 75% that the clinic would make this decision, the expected cost would be approximately $13,000. If the cost of implementing H-ABC from the point of M-ABC would cost less than $13,000, then it would be an appropriate decision to do so. The clinic did gain substantial intangible benefits by continuing to pursue the refinement of costs under the H-ABC costing method. By enlisting personnel from many areas in the cost refinement process, they increased the awareness around the clinic of the importance of cost control. Additionally, they proved the HD treatment to be profitable, which coincides with their intuition and enables them to confidently keep the treatment as an option.

Question 3

Accountants are mentored in the concepts of GAAP from the beginning of their education in accounting. Dr. Borthick contends that GAAP is a major stumbling block in the path of accountants who would like to implement flexible systems. She speaks of the need for accountants and information system specialists to make the transition from the transaction based systems of the past to new systems. These new systems would allow users to ask questions and receive answers in the format they are seeking rather than in a format prescribed by system creators. Based on our class discussions, we learned that the relational database model may provide the answer to the needs of users who are crying out for real-time information at their fingertips. With the relational model, the data is logically stored in many tables. Under the REA design, these tables are in the form of resources, events and agents. Certainly to the accountant, who was reared in the traditional file system approach this tabular approach may create some skepticism. Probably the first question from the auditor would be something similar to, "Where are the accounts receivable and accounts payable files". The auditor must be convinced that the accounts receivable and accounts payable information actually exists within the ABC system as reliable data. This is why Dr. Borthick considers GAAP a deterrent to the implementation of a flexible system, because the relational models that are necessary for flexibility and ease of information retrieval appear to conflict with the concepts of GAAP at first glance. It is only when the accountant can look past double-entry bookkeeping and traditional account titles and realize the power of SQL in deriving the necessary GAAP information. The auditors must be shown that even though the ABC design does not have a table titled "Accounts Payable" or "Accounts Receivable", these account values are intrinsic to the database design and may be derived through relationships among entities. For example in the ABC system, the accounts receivable balance can be derived through calculations using the number of treatments for the period of concern and reducing the revenue associated with these treatments by the cash receipts associated with the period of concern. Furthermore, the auditors can be shown, through the use of slightly more complex SQL statements that a listing of accounts receivable for each patient can be derived just as quickly. The auditors will also want to know how they can verify the reliability of the data contained within such a dispersed system. A number of system controls could be explained that address the issues of data input, data access and data safeguarding. The clinic would have a number of controls in place to protect the integrity of their data. Data input controls are critical and they begin with the design of the system. The auditors would first look to the REA E-R diagram, to ensure that the cardinalities chosen coordinate with the operational policies and procedures at the clinic. Specifically the minimum cardinalities will give the auditors an indication of inputs that are required by the system. For example the patient number must be input before treatment is given to ensure that the treatment is being administered to an actual patient. The auditors can verify this on the E-R diagram by noting that for each "treatment provided" event there must be at least one patient. In addition, the auditors would want to ensure that edit checks exist within the system to catch inadvertent or fraudulent attempts to enter data that is beyond practical limits or in the wrong format. For example the clinic accounts for supplies used in the treatment process, a field check would indicate an input error if the supply custodian entered a supply number that contained an alphabetic character, if typically all supplies had numeric characters only. Also a limit check in the same situation would indicate an error if the nurse were to input a supply quantity that was beyond a reasonable per treatment limit for the particular supply. The auditors would need to check the established database subschema to ensure that data access is restricted as appropriate. Certainly a system that enabled the nurses to edit the payroll information or allowed the support staff to input supply amounts used would be a red flag to any auditor. In addition, a system of passwords should exist. Input terminal locations should be convenient, but should not be located in areas where the system could easily be breached by unauthorized personnel. Backing up the data on a routine basis would be essential to prevent data loss. Additional controls such as a secondary backup site should be considered. This would safeguard data from catastrophic loss in the event of internal or external threats occurring. The auditors would definitely want to establish how the duties were divided amongst the clinic personnel to ensure that adequate separation of duties provided the necessary internal control. In the case of the clinic, ensuring that personnel authorizing patients were not handling cash receipts would be critical. Otherwise, they could submit fictitious patients to the providers and cover it up by intercepting the reimbursements. In conclusion, even though auditors will have concerns, a properly designed ABC costing system would still satisfy the requirements of GAAP, while providing the flexibility and internal control necessary to the clinic.

Question 4

The ABC article describes how accountants can work with clinics in refining aggregated numbers into informative disaggregated numbers. These disaggregated numbers can assist clinic managers in making critical decisions regarding the continuance of particular treatments. However, the ABC article provides refined information based on modifying the costing assumptions. It does not indicate anything about changing a legacy information system or implementing a new system. The ability to make real-time decisions based on H-ABC costing would require the implementation of a new information system or modification of an older system. To allay Dr. Borthick's concerns, we would need to detail the plans for implementing such a system. Unfortunately the ABC article is an example based on a specific case, which does not provide a road map to get from today's system to the ABC system of the future. In my opinion, her concerns would be stronger than ever if she were to read this article. She would be imagining the response, if every clinic around the country were to read this article. Conversations would be overheard that sound similar to the following: "Hey did you hear about that clinic that used H-ABC costing to prove that their treatments were profitable?" "Yeah, I've already asked my accountant to get together with my computer person and design that system for my clinic." "Oh, that sounds like a good idea, I think I'll call my accountant also." The net effect, thousands of accountants and IS professionals around the country asked to either upgrade a legacy system or to completely design a new system. In other words, it would create greater pressure from users than already exists, while at the same time not supplying the IS research necessary to implement the system.

On the other hand the article would be refreshing to Dr. Borthick by possibly conveying to readers that the human element in the information gathering process is as critical as the information technology that provides the information. For instance, in moving from the original assumptions under RCC costing, the staff concerns about the adequacy of those assumptions was a necessary factor. Furthermore, when M-ABC costing implied that the HD treatments were unprofitable, it was the staffs intuition that it was not a loser that motivated the more refined assumptions under H-ABC costing. Important conclusions were arrived at through the collaboration of medical staff and accountants. Dr. Borthick would certainly look at this cooperative effort as a step towards information providers and information users fulfilling each others needs. In conclusion, what would make Dr. Borthick most pleased would be, if another paper was written that detailed the implementation of an REA model relational database, which would provide real-time H-ABC cost data to dialysis clinics. This would allow all clinic information users to have their information how and when they want it.

Question 5

Event based REA model information systems and transaction based information systems are similar in that they should both arrive at the same aggregated financial statement numbers. However, the similarities may end there. The critical difference between the models is their information producing flexibility. REA models enable each user to relate data with resources, events and agents. This creates the freedom to conduct an infinite number of queries using non-referenced data. On the other hand, transaction based systems contain data that is preliminarily referenced to "best choice" categories. This creates inflexibility, since this method of data storage does not lend itself to disaggregating. It is this inflexibility that is described in the second quote. Under the REA model great care is not required to pre-select the categories for aggregation, since it is the user who has the flexibility to aggregate and disaggregate the data as they find most relevant. The first quote raises an important issue, that under the REA model we can take aggregated information, break it down to its elements and then reaggregate it, which may lead to different conclusions. The fundamental tenant is that upon reclassifying costs into subclasses the sum of the parts must equal the original aggregate amount. This implies that if the assumptions used to generate the subclasses are not sound, the redistribution of costs will merely create apparent winners and losers. Therefore the inflexibility of the transaction model may have some benefits in that it does reduce the risk of arbitrary cost allocation. In conclusion, when logical assumptions are made the refinement of information that can be provided to decision-makers under the REA model can give them a competitive advantage as in the case of H-ABC costing. However, ill-advised assumptions can provide misleading cost allocations that can result in inappropriate decision-making as in the case of M-ABC costing.