Categorizing Patients from Medical Claims Data - the Influence of GIGO

To the Editor: A substantial portion of the medical literature, particularly in managed care, comprises research using administrative claims data. Nearly every issue of the Journal of Managed Care Pharmacy has at least one article on cost or other outcomes derived from research using administrative claims data, and this issue is no exception. Caution should be exercised when evaluating the results of studies using administrative claims data. Some examples from the front lines of neurology may be helpful to understand this point. When examining opioid abuse from medical claims data for example, the first issue is the accuracy of the categorization. In the “real world” of patient care, the use of specific ICD-9-CM codes depends upon the feedback that will be given. For example, in the early years of the use of codes, physicians learned not to code patient visits for “tension headache” because the diagnosis for migraine was in the “organic disease” section while the diagnosis for tension headaches was in the “psychiatric” section. Any use of a tension headache diagnosis would invite an insurance company to reduce its payments for the patient services (psychiatric illness was paid at a lower rate than structural disease). So, the outcome of coding “accurately” for tension headache often resulted in an angry patient and lower payment rates for the physician. Miraculously, everyone seemed to have migraine. At a very fundamental level, one should not believe codes that attempt to separate patients into “abuse” versus “no abuse” unless you know very clearly who is doing the coding. It is important to know the feedback systems the coder will encounter for the coding produced. If an “abuse” code will lead the coder (doctor or other) to reprisal abuse (hostile patient, reduced payment, etc.), you can count on few accurate codes in medical claims. An even greater issue is “abuse according to what criteria?” I guarantee that patients who come to the office seeking opioids for reported pain do not tell us that they are selling the drugs, or that they really don't have pain, or that they are treating anguish rather than nociception, or the like. All of the patients report that they are miserable and that they need the medication to quell the purported illness. If determinations of abuse/nonabuse are founded simply on patient claims, then abuse will be very rare indeed. Currently, in my own practice, I have several patients who could be labeled as “abusing” opioids. That is, they are using them excessively and, in my opinion, using them really to treat anguish— the font of addiction behavior—more than nociception. One of these people is a 17-year-old grossly depressed boy who is emotionally miserable. He has tried to commit suicide. With constant effort, he is finally seeing a psychiatrist. Over time, he is given narcotics for claims of pain labeled as headache. I have no independent biological tool that will confirm or deny that he actually has pain. But, his story is “correct,” and this is the foundation I am obligated to use, based on normal standards of patient care. It is troubling that he seems to find relief in nothing but opioids, though many therapies have been tried. So, I’ve been concerned about what was really being treated. However, he was in my office, complaining of pain, miserable, depressed, and needing guidance toward relief. If you were his physician, and knew he was probably using narcotics to treat anguish more than nociception, would you say “You’re a drug abuser. Get out of my office!” How therapeutic would that be for this lost soul? Another “abuser” is a 65-year-old, extremely lonely and depressed woman going through a divorce. Similarly, for her “headache,” nothing seems to work but opioids. Similarly, she is finally seeing a psychiatrist, but only after much effort. Similarly, all early efforts to find a more constructive approach with preventatives, abortives, and other measures seem to fail. She seemed to find solace nowhere but cowering in a pill bottle, tragically trying to protect herself from her existential angst. She clearly used too many opioids. But, should we cast her off to condemnation as an abuser? How does a physician respond when she comes to the office reporting headaches? The issues are not simple. Life is not simple. There are a host of people suffering to the cores of their souls from existential anguish. There are those for whom a technocracy is too demanding. There are those for whom lost relationships are too devastating. There are those for whom failed careers are too demeaning. Yet, people find relief from these feelings in narcotics. We should, optimally, learn something from this. We should be more interested in the pharmacology of anguish, and anguish relief. However, we often tend to focus on judgment rather than understanding. Part of the reason for this is that narcotics (opioids) work appropriately when used for nociception; but, they tend to lead to dependence and addiction behavior when used to assuage existential anguish. So, there is reason for restraint in use. On the other hand, the bigger lessons come from understanding how and why opioids relieve anguish. We know that hiding in opioids doesn't work. But, the point is that labeling “abuse” becomes harder to define when you consider all the issues. Therefore, the attempt to evaluate the prevalence of abuse is a more difficult pursuit than a surface appraisal would suggest. Coincident with the advent of very expensive opioids, interested parties have promulgated the notion in America that pain is undertreated. Huge industries have sprung up to carry the banner of the pain patient as an underserved individual. These industries may not be interested in the notion of abuse. In the last decade, America has blossomed as a place advocating that abuse isn’t the issue; rather, the issue is undertreatment. Of course, predictably, a backlash would eventually happen. And, that is our current era. LETTERS

headache. I have no independent biological tool that will confirm or deny that he actually has pain. But, his story is "correct," and this is the foundation I am obligated to use, based on normal standards of patient care. It is troubling that he seems to find relief in nothing but opioids, though many therapies have been tried. So, I've been concerned about what was really being treated. However, he was in my office, complaining of pain, miserable, depressed, and needing guidance toward relief. If you were his physician, and knew he was probably using narcotics to treat anguish more than nociception, would you say "You're a drug abuser. Get out of my office!" How therapeutic would that be for this lost soul?
Another "abuser" is a 65-year-old, extremely lonely and depressed woman going through a divorce. Similarly, for her "headache," nothing seems to work but opioids. Similarly, she is finally seeing a psychiatrist, but only after much effort. Similarly, all early efforts to find a more constructive approach with preventatives, abortives, and other measures seem to fail. She seemed to find solace nowhere but cowering in a pill bottle, tragically trying to protect herself from her existential angst. She clearly used too many opioids. But, should we cast her off to condemnation as an abuser? How does a physician respond when she comes to the office reporting headaches?
The issues are not simple. Life is not simple. There are a host of people suffering to the cores of their souls from existential anguish. There are those for whom a technocracy is too demanding. There are those for whom lost relationships are too devastating. There are those for whom failed careers are too demeaning. Yet, people find relief from these feelings in narcotics. We should, optimally, learn something from this. We should be more interested in the pharmacology of anguish, and anguish relief. However, we often tend to focus on judgment rather than understanding. Part of the reason for this is that narcotics (opioids) work appropriately when used for nociception; but, they tend to lead to dependence and addiction behavior when used to assuage existential anguish. So, there is reason for restraint in use. On the other hand, the bigger lessons come from understanding how and why opioids relieve anguish.
We know that hiding in opioids doesn't work. But, the point is that labeling "abuse" becomes harder to define when you consider all the issues. Therefore, the attempt to evaluate the prevalence of abuse is a more difficult pursuit than a surface appraisal would suggest.
Coincident with the advent of very expensive opioids, interested parties have promulgated the notion in America that pain is undertreated. Huge industries have sprung up to carry the banner of the pain patient as an underserved individual. These industries may not be interested in the notion of abuse. In the last decade, America has blossomed as a place advocating that abuse isn't the issue; rather, the issue is undertreatment. Of course, predictably, a backlash would eventually happen. And, that is our current era.
So, whatever study we would do to investigate the prevalence of drug abuse must begin by looking very carefully at the methods for gathering data. What influences will determine how coders apply codes? And, how shall we view the state of the patient who "overuses" narcotics? Without careful attention to these issues, the result of a study, particularly a retrospective study of administrative claims data, may be GIGO-garbage in, garbage out.

ss Guiding Principles for Effective Electronic Messaging
To the Editor: Online capabilities have greatly improved efficiency in prescription claims processing. Pharmacies receive information promptly about patient eligibility, copayments, benefits, and pharmacy reimbursement. This improved efficiency benefits all stakeholders. In addition, computerization has allowed the implementation of concurrent drug utilization review (DUR). An integrated DUR program can benefit patients by detecting potential adverse drug-drug interactions, curbing drug misuse and abuse, and monitoring quality of care.
In the past few years, however, there has been a dramatic increase in prescription volume coupled with the increasing complexity of pharmacy benefits. These 2 factors, among others, have led to increased administrative burdens on pharmacists during the prescription dispensing process. These increased administrative demands have had an impact on the time a pharmacist is able to spend on patient care services.
To address these issues, the Academy of Managed Care Pharmacy (AMCP) called together national organizations representing health care professionals, health plans, and pharmacy benefit managers to form a work group to find methods to enhance patient care and patient safety through effective electronic messaging. The task placed before the work group was to find a way to make the electronic messaging that occurs during prescription drug claims processing clearer and more helpful. The work group set a goal to achieve consensus on a set of guiding principles that outline the essential components of clear, effective, and actionable electronic messaging. The resulting document has been produced for use by all parties involved in prescription claims processing. The Guiding Principles for Effective Electronic Messaging is a tool for health system administrators, pharmacy benefit managers (PBMS), pharmacy claims processors, pharmacists, and patients.
The work group urges adoption of these voluntary principles because their use can improve communications between pharmacies, claims processors, and patients and ultimately lead to improved patient care.

Introduction to the Guiding Principles
The Work Group enumerated the Guiding Principles, recognizing that improved patient care will result if electronic messages are optimally developed and transmitted. For electronic messaging to be effective, pharmacists must be able to understand the reason the message was sent, interpret the meaning of that message easily, and perform some action based on that interpretation. A glossary and notes and references section are included to clarify terminology and to provide additional resources.
The Work Group believes that these voluntary Guiding Principles will be a valuable tool for health care system administrators, PBMS, pharmacy claims processors, and pharmacists. Each of these stakeholders plays a role in ensuring that electronic messages convey appropriate and understandable information, thus ensuring that patients have access to rational, clinically appropriate, safe, and cost-effective drug therapy.
The purpose of this document is to encourage quality patient care by improving communications between pharmacies, claims processors, patients and all other involved parties.

Pharmacy Claims Transmission Standards
Electronic processing of prescription drug claims has been used for more than 2 decades, but it would not have been possible without the development and implementation of transmission standards. In order to convert a manual process to an electronic one, all the information necessary to complete the dispensing process must be codified and standardized. The National Council for Prescription Drug Programs (NCPDP) has developed the standard code set for prescription drug product claims transmissions. NCPDP is an American National Standards Institute (ANSI)-accredited standards development organization consisting of members who represent all parties interested in electronic standardization within the pharmacy Letters services sector of the health care industry. 1 When used in this document, the words Standard v5.1 refer to the NCPDP Telecommunication Standard v5.1. The standard is composed of numerous fields containing data elements that identify the patient, drug, prescriber, pharmacy, and other items. Real-time, 2-way communication between pharmacies and claims processors is enabled by use of the standard. Its use greatly enhances the abilities of both pharmacists and processors in the provision of drug therapy.

Call to Action
The most important message of the Guiding Principles is that stakeholders involved in the provision of pharmacy services must communicate with each other clearly and effectively. Only with continual 2-way communication and the cooperation of all parties can meaningful improvements in electronic messaging be made. The Work Group believes that use of the Guiding Principles will lead to increased quality of care and patient safety. The publication of the Guiding Principles is one of many steps to improve safety and quality. The Work Group encourages the adoption and implementation of these voluntary principles by health care system administrators, PBMS, pharmacy claims processors, and pharmacists as the next step in this ongoing improvement process.
Principle I 2 : Systems should be designed to adhere to the current recognized standard. Explanation: Both pharmacy practice systems and claims processing systems should be designed using the currently recognized telecommunications standard. As of October 2003, the Health Insurance Portability and Accountability Act of 1996 (HIPAA) has mandated the standard. For prescription drug product claims, it is the NCPDP Telecommunication Standard v5.1.
Systems that are designed around Standard v5.1 ensure consistency in the implementation and use of pharmacy claims transmissions. Claims processing systems should send only standard messages in the proper fields, and pharmacy practice systems should correctly interpret and appropriately display those standard messages.
Principle II: All electronic data fields should be used only for their intended purposes as defined in the current transmission standard. Explanation: Anecdotal reports indicate that certain telecommunications data fields are sometimes used for purposes other than those defined in Standard v5.1. This practice may lead to inconsistencies and confusion in claims processing. Just as systems should be designed to adhere to the current recognized standard, users of those systems should implement those standards only as defined. NCPDP provides an implementation guide for Standard v5.1 for its members, and this document explains the purpose for and use of each data field. 3 Ezetimibe: The pharmacokinetics of ezetimibe in adolescents (10 to 18 years) have been shown to be similar to that in adults. Treatment experience with ezetimibe in the pediatric population is limited to 4 patients (9 to 17 years) with homozygous sitosterolemia and 5 patients (11 to 17 years) with HoFH. Treatment with ezetimibe in children (<10 years) is not recommended. Simvastatin: Safety and effectiveness of simvastatin in patients 10-17 years of age with heterozygous familial hypercholesterolemia have been evaluated in a controlled clinical trial in adolescent boys and in girls who were at least 1 year post-menarche. Patients treated with simvastatin had an adverse experience profile generally similar to that of patients treated with placebo. Doses >40 mg have not been studied in this population. In this limited controlled study, there was no detectable effect on growth or sexual maturation in the adolescent boys or girls, or any effect on menstrual cycle length in girls. Adolescent females should be counseled on appropriate contraceptive methods while on therapy with simvastatin (see CONTRAINDICATIONS and PRECAUTIONS, Pregnancy). Simvastatin has not been studied in patients younger than 10 years of age, nor in pre-menarchal girls.

Geriatric Use
Of the patients who received VYTORIN™ (ezetimibe/simvastatin) in clinical studies, 792 were 65 and older (this included 176 who were 75 and older). The safety of VYTORIN was similar between these patients and younger patients. Greater sensitivity of some older individuals cannot be ruled out. (See CLINICAL PHARMACOLOGY, Special Populations and ADVERSE REACTIONS.) ADVERSE REACTIONS VYTORIN has been evaluated for safety in more than 3800 patients in clinical trials. VYTORIN was generally well tolerated. Table 1 summarizes the frequency of clinical adverse experiences reported in ≥ 2% of patients treated with VYTORIN (n=1236) and at an incidence greater than placebo regardless of causality assessment from 3 similarly designed, placebo-controlled trials. Ezetimibe: Other adverse experiences reported with ezetimibe in placebo-controlled studies, regardless of causality assessment: Body as a whole -general disorders: fatigue; Gastrointestinal system disorders: abdominal pain, diarrhea; Infection and infestations: infection viral, pharyngitis, sinusitis; Musculoskeletal system disorders: arthralgia, back pain; Respiratory system disorders: coughing.

Post-marketing Experience
The following adverse reactions have been reported in post-marketing experience, regardless of causality assessment: Hypersensitivity reactions, including angioedema and rash; pancreatitis; nausea; cholelithiasis; cholecystitis. Simvastatin: Other adverse experiences reported with simvastatin in placebo-controlled clinical studies, regardless of causality assessment: Body as a whole -general disorders: asthenia; Eye disorders: cataract; Gastrointestinal system disorders: abdominal pain, constipation, diarrhea, dyspepsia, flatulence, nausea; Skin and subcutaneous tissue disorders: eczema, pruritus, rash. The following effects have been reported with other HMG-CoA reductase inhibitors. Not all the effects listed below have necessarily been associated with simvastatin therapy. Musculoskeletal system disorders: muscle cramps, myalgia, myopathy, rhabdomyolysis, arthralgias. Nervous system disorders: dysfunction of certain cranial nerves (including alteration of taste, impairment of extra-ocular movement, facial paresis), tremor, dizziness, memory loss, paresthesia, peripheral neuropathy, peripheral nerve palsy, psychic disturbances. Ear and labyrinth disorders: vertigo. Psychiatric disorders: anxiety, insomnia, depression, loss of libido. Hypersensitivity Reactions: An apparent hypersensitivity syndrome has been reported rarely which has included 1 or more of the following features: anaphylaxis, angioedema, lupus erythematous-like syndrome, polymyalgia rheumatica, dermatomyositis, vasculitis, purpura, thrombocytopenia, leukopenia, hemolytic anemia, positive ANA, ESR increase, eosinophilia, arthritis, arthralgia, urticaria, asthenia, photosensitivity, fever, chills, flushing, malaise, dyspnea, toxic epidermal necrolysis, erythema multiforme, including Stevens-Johnson syndrome. Gastrointestinal system disorders: pancreatitis, vomiting. Hepatobiliary disorders: hepatitis, including chronic active hepatitis, cholestatic jaundice, fatty change in liver, and, rarely, cirrhosis, fulminant hepatic necrosis, and hepatoma. Metabolism and nutrition disorders: anorexia. Skin and subcutaneous tissue disorders: alopecia, pruritus. A variety of skin changes (eg, nodules, discoloration, dryness of skin/mucous membranes, changes to hair/nails) have been reported. Reproductive system and breast disorders: gynecomastia, erectile dysfunction. Eye disorders: progression of cataracts (lens opacities), ophthalmoplegia. Laboratory Abnormalities: elevated transaminases, alkaline phosphatase, γ-glutamyl transpeptidase, and bilirubin; thyroid function abnormalities.

Laboratory Tests
Marked persistent increases of serum transaminases have been noted (see WARNINGS, Liver Enzymes). About 5% of patients taking simvastatin had elevations of CK levels of 3 or more times the normal value on 1 or more occasions. This was attributable to the noncardiac fraction of CK. Muscle pain or dysfunction usually was not reported (see WARNINGS, Myopathy/Rhabdomyolysis).

Concomitant Lipid-Lowering Therapy
In controlled clinical studies in which simvastatin was administered concomitantly with cholestyramine, no adverse reactions peculiar to this concomitant treatment were observed. The adverse reactions that occurred were limited to those reported previously with simvastatin or cholestyramine. Adolescent Patients (ages 10-17 years) In a 48-week controlled study in adolescent boys and girls who were at least 1 year post-menarche, 10-17 years of age with heterozygous familial hypercholesterolemia (n=175), the safety and tolerability profile of the group treated with simvastatin (10-40 mg daily) was generally similar to that of the group treated with placebo, with the most common adverse experiences observed in both groups being upper respiratory infection, headache, abdominal pain, and nausea (see CLINICAL PHARMACOLOGY, Special Populations and PRECAUTIONS, Pediatric Use). Example: The definition of Reject Code 70 is "Product/Service not covered;" in contrast, the definition of Code 61 is "Product/Service not covered for patient gender." The second code is more specific and conveys additional important information that can be used by the pharmacist to explain the coverage rejection to the patient and/or determine a moreappropriate medication to recommend to a physician. The simpler code definition "Product/Service not covered" does not provide enough information for the pharmacist to take action. In this example, Reject Code 61 should not require any free-text message explanation-the definition of this code provides enough information for the pharmacist to take action.
Principle IV: Additional textual information may have to accompany the standard message code to make it actionable. Explanation: Providing additional information to supplement the standard reject code can help to clarify the message and suggest actions that the pharmacist may perform. That information must be transmitted in the correct field(s) as described in the currently recognized telecommunications standard.

Example:
The definition of Reject Code 60 is "Product/Service is not covered for patient age." Additional information such as the minimum or maximum patient age allowed for coverage of the product/service allows the pharmacist to easily interpret the meaning of the message and allows him/her to take an action.
Principle V: Systems should provide the pharmacist with all significant information during the processing of prescription drug claims.

Explanation:
In an effort to reduce redundancy, both parties to a claims transmission need to fully inform the other. Each time a prescription is dispensed, there are at least 2 systems in use: pharmacy practice systems and claims processing systems. Most pharmacy practice systems screen for the same drug-drug interactions as the claims processors' systems. If both systems identify the same interaction, redundant messages may be transmitted to the pharmacist. In an attempt to reduce or eliminate these duplicate messages, some pharmacy systems suppress the online messages from the claims processor. This practice may impact patient care, especially in instances where patients use multiple pharmacies to fill prescriptions and claims processors' systems contain information that individual pharmacy systems do not. Pharmacy systems, which may have additional information (e.g., over-the-counter products, herbal remedies), should be designed in such a manner that all significant transmitted information is displayed to pharmacy personnel without the redundancy that can be caused by the 2 systems.
Principle VI: Pharmacy systems should transmit with a claim any relevant information about their actions on a claim.

Explanation:
In an effort to reduce redundancy, both parties to a claims transmission need to fully inform the other. While filling prescriptions, the pharmacist often identifies drug-drug interactions or other DUR alerts that in his/her professional judgment are considered minor and should not affect the health of the patient. After consultation with the prescriber, many of these prescriptions are filled. Yet, the claims processor will most likely identify and flag the same interaction, the result of which is the transmission of a DUR alert message to the pharmacist.
Pharmacists and their pharmacy practice systems should use the Professional Pharmacy Service (PPS) fields that exist in Standard v5.1 to inform the claims processor that they have already identified the interaction and taken action. The processor can use this information to prevent the transmission of those same drugdrug interaction messages. 4 This does not preclude the processor from sending all information deemed necessary.

Principle VII: Processors should not transmit information that is reasonably determined to be redundant.
Explanation: In most instances, the pharmacist benefits from having more information; however, there are standard reject codes that are understandable and actionable without the addition of text. In those instances, the supplemental data may be viewed as redundant and "noisy." Example: Transmitting a message that simply restates the definition of the reject code is redundant, provides no value, and may be viewed by the pharmacist as "noise." As stated in Principle III, Reject Code 61 ("Product/Service not covered for patient gender") should not require additional explanation; therefore the transmission of Reject Code 61 with an auxiliary message such as "Drug not covered for females" is not necessary.

Principle VIII: The use of abbreviations in free-text messages is strongly discouraged.
Explanation: Each sector of the medical community (pharmacists, health plans, prescribers, etc.) has its own set of abbreviations. Many abbreviations may have more than one meaning, based on the health care sector and use. Because of the potential for misinterpretation, patient safety concerns dictate that their use in free-text fields should be discouraged. With the implementation of Standard v5.1, the size of the field used for supplemental text messages has been increased to 200 characters. This should provide ample space to convey messages without the use of abbreviations.

Letters
Glossary 5 American National Standards Institute (ANSI). The American National Standards Institute is a private, nonprofit organization that administers and coordinates the U.S. voluntary standardization and conformity assessment system. The Institute' s mission is to enhance both the global competitiveness of U.S. business and the U.S. quality of life by promoting and facilitating voluntary consensus standards and conformity assessment systems and safeguarding their integrity.
Claim. The information submitted by a provider or a covered person to establish that health care services were provided to a covered person. The claim is used for processing payment to the provider or covered person.
Claims processing/adjudication. Applying the terms of a health care insurance policy or public benefits to determine eligibility, coverage, proper payments, and/or share of responsibility for payments between insurer and covered person.
Claims processor. A company, on behalf of a payer, that applies eligibility and medical insurance benefit rules in determining the reimbursement/payment amount of services provided by health professionals.
Coverage/claim rejection. Claims that are denied payment by a claims processor.

Drug utilization review (DUR).
A system of drug use review that can detect potential adverse drug interactions, drugpregnancy conflicts, therapeutic duplication, drug-age conflicts, etc. There are 3 forms of DUR: prospective (before prescription dispensing), concurrent (at the time of prescription dispensing) and retrospective (after the therapy has been completed). Appropriate use of an integrated DUR program can curb drug misuse and abuse and monitor quality of care. DUR can reduce hospitalization and other costs related to inappropriate drug use.
Electronic messaging. The electronic exchange (through computers) of information between 2 or more organizations using standardized formats, code sets, and guidelines/rules.
Health care system administrator. The person responsible for configuring, administering, and maintaining the claims processing systems within a health plan/pharmacy benefit manager (PBM).
Health Insurance Portability and Accountability Act of 1996 (HIPAA). The legislation intended to provide portability of employer-sponsored insurance from one job to another in order to prevent what has become known as "job lock," or the inability to change jobs because of the fear of losing health insurance. The legislation also includes regulations for maintaining the pri-vacy of health information (privacy regulations) and specifies the manner in which that information can be transmitted electronically (transmission and code sets regulations).
Health plan. Organizations such as a health maintenance organization (HMO), a preferred provider organization (PPO), an insured plan, a self-funded plan, or another entity that covers health care services.

National Council for Prescription Drug Programs (NCPDP).
An organization created to promote data interchange standards for the pharmacy services sector of the health care industry. Membership is open to all segments of the prescription drug industry.
Patient care services. The actions or products provided by a health care professional to another person to provide some affect on the health of that person.

Pharmacy benefit manager (PBM).
An organization that manages pharmaceutical benefits for a managed care organization, other medical providers, or an employer. PBMs contract with clients interested in optimizing the clinical and economic performance of their pharmacy benefit. PBM activities may include some or all of the following: benefit plan design, creation/administration of retail and mail-service networks, claims processing and managed prescription drug care services such as drug utilization review, formulary management, generic dispensing, prior authorization, and disease and health management.
Pharmacy systems. The computer software and hardware that address the business needs of the pharmacy and claims processor.
Claims processing system. The computer system designed to provide instant online adjudication of third-party prescription drug claims at the point of service.
Pharmacy practice system. The computer system that helps pharmacists keep track of day-to-day pharmacy operations. The system handles all aspects of prescription processing, including patient information, drug information, and third-party billing.
Prescription dispensing process. The act, as defined by state laws, of a pharmacist in the proper distribution of prescription drugs to a patient pursuant to a legal order from a prescriber.
Reject code. One of a defined set of values within a standardized list that always translates to the same meaning.
Transmission standards. The electronic transfer of information such as electronic media health claims, in a standard format. They provide standardization of the data content by specifying uniform definitions of the data elements that are exchanged and the code Letters sets used in those transactions. A code set refers to any set of codes used to encode data elements, such as tables of terms, medical diagnostic codes, or drug/product identifiers, etc.

NOTES AND REFERENCES
1. Additional information about NCPDP and pharmacy claims transmission standards may be found on its Web site: www.ncpdp.org.
2. The Guiding Principles are numbered for clarity only. The numbering does not denote a hierarchy or designate a sequence for implementation.
4. The intent of Principle VI is to reduce the number of duplicative and unnecessary messages. Some parties believe that transmitting information to the claims processor about actions taken by a pharmacist prior to the filling of a prescription exceeds the HIPAA privacy rule (all parties are required to make reasonable efforts to limit the use, disclosure of, and requests for protected health information to the minimum necessary to accomplish the intended purpose). Given that the Office for Civil Rights has not issued a ruling on the practice described in Principle VI, the Work Group decided it should remain in the document. Additional state laws may be applicable and should be considered.

WORK GROUP PARTICIPATING ORGANIZATIONS
AMCP would like to thank the following organizations that provided input and comments in the writing of this document. The organizations listed below should not be considered as endorsers of the content but rather contributors of information contained within the document.

ss Dose Consolidation Can Be an Efficient Intervention
To the Editor: Delate, Fairman, Carey, and Motheral in the September/ October 2004 issue of JMCP reported unsatisfactory results associated with a letter-based dose consolidation program aimed at decreasing pharmacy expenditures. 1 Their conclusion that a letter-based dose consolidation program may not significantly decrease pharmacy expenditures should be interpreted cautiously as this finding has not been supported by other studies, and their results may be attributable, at least in part, to their specific program design and practice setting. Their findings were contradictory to the savings we have previously reported from the dose consolidation program in our organization. 2 Differences in their program, compared with ours, may explain the disparity in results. Understanding these differences is critical to anyone involved in the design and implementation of a successful dose consolidation program. We published the results of a dose consolidation program using our proprietary, Web-based therapeutic intervention application. Our program was conducted in an "at-risk" physician environment supported by local pharmacy staff using academic detailing. Our system generates preformatted authorization forms, preprinted prescriptions for the suggested alternative dosage regimen, and personalized letters from the clinician to the patient for each dose consolidation intervention. These materials are presented to physicians during one-on-one and group educational sessions where the physicians are asked to review the patient-specific request for conversion, indicate whether it is clinically appropriate, and sign the patient letter and new prescription if they are in agreement with the change. All completed interventions are processed by our clinical pharmacy staff and tracked via our Web-based therapeutic intervention application.
Delate et al. reported results from a trial conducted in a large health plan in which potential interventions were randomized to either a letter to the physician-only arm or a letter to the physician and patient arm. It is not surprising to us that sending a letter to a physician, not "at-risk" for pharmacy costs, without a recent drug profile, new prescription, or patient letter to sign, results in poor physician compliance with the requested intervention. The differences in design between their program and ours likely explain the varying results. Looking at their experience and ours together, one may conclude that a pharmacist-facilitated program, with appropriate supports for the physician, can deliver results that a pure mail-based outreach program, without significant supports for the physician, cannot.
The criteria for identifying opportunities for intervention also varied significantly between the 2 dose consolidation programs. Delate et al. utilized a ratio of metric quantity-todays supply of 1.5 to 2.4 to identify patients suitable for dose consolidation, whereas our program utilized a whole-number Letters ratio, primarily 2.0. The methodology utilized by Delate et al. likely led to overidentification of potential targets for conversion. For example, if a patient is receiving a 75 mg daily dose of sertraline (Zoloft), the ratio of monthly quantity (#45) to days supply (#30) would be 1.5, yet this patient would not be a valid candidate for conversion because no 75 mg sertraline (Zoloft) dosage form is available.
We utilized our health system data to evaluate the potential impact of using the Delate  In all likelihood, this overidentification also contributed to the lower overall success rate demonstrated in their program compared with ours.
It is also noteworthy that the study by Delate et al. used pharmacy claims to estimate the expected refill rates to account for possible noncompliance and used these figures, rather than annualized savings. This approach does not account for the fact that savings are likely to continue to accrue for a fair number of patients who stay on the "consolidated" drug regimen and with the health plan beyond 1-year postintervention.
Administrative costs were also factored into their cost analysis, something that was not done in our study because our costs, with the exception of mailing costs, were largely fixed. The pharmacy staff utilized to support our initiative was already employed with the organization and was meeting routinely with providers prior to implementing the program. Although there was an initial investment in developing our Web-based therapeutic interchange tool, this tool also provides support for several interventions, not solely dose consolidation. Accounting for ongoing development support and direct pharmacist time, this tool has consistently provided a 5:1 return-on-investment since its inception. Mailing costs for our program are less than 0.1% of the cost savings achieved.
Our therapeutic substitution program continues to generate significant savings since our initial report published in JMCP. In 2002, 32% of all dose consolidation interventions presented to clinicians were approved, yielding an estimated annualized drug cost savings of $0.05 per member per month (PMPM). In 2003, another 46% of all dose consolidation opportunities presented were approved, which accounted for approximately $0.09 PMPM in annualized drug cost savings. As we gain experience with our program, we improve our methods for targeting patient drug regimens and supporting the interventions, resulting in the increase in our percentage of opportunities that is approved by our clinicians. Our total return-on-investment from dose consolidation improved from 2002 to 2003.
In contrast to the experience of Delate et al., our dose consolidation program continues to decrease the growth rate of pharmacy expenditures within our health system. As the cost of many agents targeted for dose consolidation continues to increase at a significant rate each year, interventions to consolidate dosing remain an important element of our overall program to improve the quality and cost efficiency of prescribing within our system. We suggest that readers not lump dose consolidation programs into a single category but endeavor to determine the differences in the methods of the interventions that may explain apparently disparate results.

The Authors Respond
We thank Dr. Baldinger and Mr. Calabrese for their observations.

Letters
We agree with them that readers should consider differences in intervention methods when assessing the outcomes of any pharmacy benefit management program. We also agree that the program that they studied in 2002 1 is much more intensive than the one examined in our study and might produce different outcomes. However, a key premise of their letter, that our financial results are contradictory to theirs, is incorrect. In fact, as we point out in our report, 2 we found similar expenditure decreases ($0.11 per member per month [PMPM] dollars for the Physician/Member Letter Arm compared with the $0.15 PMPM in 2003 dollars that they reported), when using a calculation methodology similar to theirs. It is only when we incorporated empirically determined compliance rates for each therapy class and took into account the natural course of pharmacotherapy (i.e., the expenditure change seen in the control group) that we saw the expenditure decrease to $0.03 PMPM for that study arm. These findings, coupled with the rate of dose consolidation observed in our control group that received no intervention at all, highlight the inadequacy of the nonexperimental research designs that have been employed in previous dose consolidation evaluations. In assessing program cost-effectiveness, using a randomized controlled design as we did, is optimal. 3 Moreover, we question Baldinger and Calabrese' s argument that it is unnecessary to take administrative costs into account when using already-employed staff to carry out program activities. It is essential to consider all program costs-both fixed and variable-to provide a realistic assessment of a pragmatic dose consolidation program in a typical health plan.
We agree that our use of a wider ratio range to identify inefficient doses may have resulted in targeting false-positive inefficient doses. But, as we point out in our report, 2 if there was false-positive targeting of inefficient doses, any potential bias would have been randomly distributed across arms due to the study design and not had a noticeable impact on our findings.
Retrospective dose consolidation programs are a widely implemented pharmacy management strategy that have a natural appeal for use in lieu of a prospective point-of-service quantity management strategy. However, our findings indicate that when these programs are properly assessed (i.e., taking into account all administration costs, drug discontinuation, natural rates of consolidation, and patient noncompliance), these programs do not substantially decrease pharmacy expenditures for a typical health plan.