Survey of the Average Cost of Dispensing a Medicaid Prescription in the State of Texas

Transcription

1 Survey of the Average Cost of Dispensing a Medicaid Prescription in the State of Texas Prepared for the Texas Health and Human Services Commission June

2 Table of Contents CHAPTER 1: EXECUTIVE SUMMARY...4 INTRODUCTION... 4 SUMMARY OF FINDINGS...5 Table 1.1 Dispensing Cost for Texas Medicaid Pharmacies...5 CONCLUSIONS AND OPTIONS FOR A PHARMACY DISPENSING FEE...5 Table 1.2 Approximate Values for Fixed Component Portion of a Variable Dispensing Fee Structure...7 CHAPTER 2: COST OF DISPENSING METHODOLOGY AND FINDINGS...8 METHODOLOGY OF THE DISPENSING COST SURVEY...9 Survey Distribution... 9 Table 2.1 Dispensing Cost Survey Response Rate...11 Tests for Reporting Bias Review Procedures Desk Reviews On-Site Field Examinations COST FINDING PROCEDURES...12 Overhead Costs Table 2.2 Non-Allocated Expenses per Prescription...15 Labor Costs Limits on Reasonable Cost Owner Compensation Issues Overall Labor Cost Constraints Inflation Factors DISPENSING COST ANALYSIS AND FINDINGS...20 Table 2.3 Dispensing Cost per Prescription All Pharmacies...21 Specialty Pharmacies Table 2.4 Dispensing Cost per Prescription - Specialty Versus Other Pharmacies...22 Table 2.5 Dispensing Cost per Prescription Excluding Specialty Pharmacies...23 Relationship of Dispensing Cost with Prescription Volume...23 Table 2.6 Dispensing Cost by Pharmacy Total Annual Prescription Volume...24 Table 2.7 Statistics for Pharmacy Total Annual Prescription Volume...24 Other Observations Associated with Dispensing Cost and Pharmacy Attributes...24 Table 2.8 Components of Prescription Dispensing Cost

3 EXHIBITS Exhibit 1 Exhibit 2 Texas Medicaid Pharmacy Cost of Dispensing Survey Survey Form Texas Medicaid Pharmacy Cost of Dispensing Survey Instructions Exhibit 3a Instruction Letter from the Texas Health and Human Services Commission (Independent Pharmacies) Exhibit 3b Instruction Letter from the Texas Health and Human Services Commission (Chain Pharmacies) Exhibit 4a First Survey Reminder Letter (Independent Pharmacies) Exhibit 4b First Survey Reminder Letter (Chain Pharmacies) Exhibit 5a Second Survey Reminder / Extension Letter (Independent Pharmacies) Exhibit 5b Second Survey Reminder / Extension Letter (Chain Pharmacies) Exhibit 6 Exhibit 7 Exhibit 8 Exhibit 9 On-site Field Examination Notification Letter Summary of On-site Field Examination Findings Construction of Owner Pharmacist Salary Limits Table of Inflation Factors for Dispensing Cost Survey Exhibit 10 Histogram of Pharmacy Dispensing Cost Exhibit 11 Pharmacy Cost of Dispensing Survey Data - Statistical Summary Exhibit 12 Charts Relating to Pharmacy Prescription Volume: A: Histogram of Pharmacy Total Prescription Volume B: Scatter-Plot of Relationship between Dispensing Cost per Prescription and Total Prescription Volume Exhibit 13 Chart of Components of Cost of Dispensing per Prescription Exhibit 14 Summary of Pharmacy Attributes 3

4 Chapter 1: Executive Summary Introduction Under contract to the Texas Health and Human Service Commission (HHSC), Myers and Stauffer LC performed an analysis of the cost that pharmacies incur to provide services to Medicaid beneficiaries. This analysis included both a study of pharmacy dispensing cost and an evaluation of options for an ingredient reimbursement methodology. This report includes information regarding the methodology and findings of the pharmacy cost of dispensing survey. A separate report from Myers and Stauffer includes an analysis of options for a pharmacy ingredient reimbursement methodology. The pharmacy cost of dispensing study followed the methodology and used a survey instrument similar to those used by Myers and Stauffer in Medicaid pharmacy engagements in several other states. The methodology was consistent with guidelines from the Centers for Medicare and Medicaid Services (CMS) regarding the components of pharmacy cost that are appropriately reimbursed by the pharmacy dispensing fee of a state Medicaid program. Myers and Stauffer obtained from the HHSC a list of pharmacy providers currently enrolled in the Texas Medicaid pharmacy program. According to the provider list, there were 4,433 pharmacy providers enrolled in the program that were located within the state of Texas. Myers and Stauffer used a random sampling approach to develop a sample of 2,000 pharmacy providers which were to be included in the pharmacy cost of dispensing survey. These 2,000 pharmacies were requested to submit financial and other relevant information for this study. Myers and Stauffer performed desk review procedures to test completeness and accuracy of all dispensing cost surveys submitted. Additionally, on-site field examinations were performed at 32 pharmacies to validate reported costs. There were 1,263 pharmacies that filed cost surveys that could be included in this analysis. Data from these surveys in conjunction with cost-finding algorithms were used to calculate the average cost of dispensing at each pharmacy and these results were subjected to statistical analysis to determine statewide averages of the cost of dispensing. 4

5 Summary of Findings Based on the results of the survey, the median cost of dispensing, weighted by Medicaid volume, was $9.47 per prescription for all pharmacies including specialty pharmacies. For non-specialty pharmacies only, the mean cost of dispensing, weighted by Medicaid volume, was $8.98 per prescription. Table 1.1 reports additional values that measure the average cost of dispensing determined from the survey data. Table 1.1 Dispensing Cost for Texas Medicaid Pharmacies All Pharmacies Inclusive of Specialty Nonspecialty Pharmacies Only Pharmacies Included in Analysis 1,263 1,206 Weighted Median A, B $9.47 $8.98 Weighted Mean (Average) A,B $11.57 $10.12 A Inflated to common point of June 30, 2013 (midpoint of state fiscal year ending December 31, 2013). B Weighted by Medicaid volume. Conclusions and Options for a Pharmacy Dispensing Fee There are several factors that should be considered in determining appropriate Medicaid pharmacy dispensing fees. These include: Cost incurred by pharmacies to dispense and acquire prescriptions for Medicaid recipients. Market dynamics, including an understanding of the payment rates accepted by pharmacies from other payers. The need to maintain sufficient patient access to pharmacy services for Medicaid recipients throughout the state. Currently, the Texas Medicaid program uses a pharmacy dispensing fee that includes both a fixed component and a variable component. The fixed component of the dispensing fee is $6.50 and the variable component applies a factor of to the ingredient cost for the drug dispensed. Payment for a prescription medication is determined by adding $6.50 (the fixed component) to the estimated acquisition cost (EAC) of the drug and dividing the sum by (the variable component). Additional adjustments to the dispensing fee are made based on the provision of delivery services and the dispensing of prescriptions designated as preferred generic products. The total dispensing fee is capped at $200. 5

6 Changes in the Medicaid pharmacy reimbursement formula should consider both the dispensing and ingredient components of the payment structure. And should take into consideration the proposed rule regarding Medicaid pharmacy services published by CMS on February 2, This proposed rule, if finalized in its present state, would change the current basis for ingredient reimbursement from the currently defined estimated acquisition cost (EAC) with the concept of actual acquisition cost (AAC). Further analysis of options for pharmacy ingredient reimbursement have been presented to the HHSC by Myers and Stauffer in a separate report. In conjunction with the requirement for ingredient reimbursement based on the AAC, the proposed rule from CMS has reiterated the importance of the pharmacy dispensing fee. CMS has indicated that state Medicaid programs that make the switch to an AAC methodology will be required to also implement a professional dispensing fee that reflects the pharmacist s professional services and costs associated with the dispensing of drug products to Medicaid recipients. One option available to the HHSC is to use a single statewide pharmacy dispensing fee which would be the same amount for all providers regardless of the prescription being dispensed. This is the most common approach for dispensing fees used by state Medicaid programs. Based on the results of the study of pharmacy dispensing cost, a single statewide dispensing fee of $8.98 would reimburse the weighted median cost of dispensing prescriptions to Medicaid recipients for non-specialty prescriptions. Alternatively, were the HHSC to retain a variable dispensing fee structure as is currently used, an adjustment to the fixed component of the dispensing fee would be necessary to set the variable dispensing fee such that the average dispensing fee paid is equivalent to the average cost of dispensing as observed by the cost of dispensing survey. Myers and Stauffer used a sample of Texas Medicaid pharmacy claims data to model the approximate impact of a variable dispensing fee structure based on revisions to the ingredient cost basis to more closely reflect average acquisition cost. The table below indicates the approximate fixed component that would be needed to realize average dispensing fees at levels that reflect the results from the cost of dispensing survey. For each of these scenarios, the retention of the current variable component of the dispensing fee of is assumed. Since the HHSC has not yet finalized its selection of an ingredient reimbursement methodology, this analysis of the components of a variable dispensing fee is preliminary. Further analysis of these parameters should be performed after the HHSC has determined all of the details of a new ingredient reimbursement methodology. Depending on the timeframe in which the HHSC were to implement a dispensing fee based on the results of the cost of dispensing survey, consideration of the application of an inflation factor may be appropriate to account for the time passed since the fiscal reporting cycles of the financial data that formed the basis of the survey. Myers and Stauffer used the Employment Cost Index (ECI) published by the Bureau of Labor Statistics to adjust the dispensing cost data to the common point of June 30, 2013 (i.e., the midpoint of a fiscal period ending December 31, 2013). Recent trends in the ECI reflect an annual inflation trend of approximately 1.9% on an annual basis. This trend could be used to further adjust the dispensing cost results from the survey as needed for setting future dispensing fees. 6

7 Table 1.2 Approximate Values for Fixed Component Portion of a Variable Dispensing Fee Structure Measure of Average Cost of Dispensing Median Weighted by Medicaid Volume for all Pharmacies (excluding specialty pharmacies) Mean Weighted by Medicaid Volume for all Pharmacies (excluding specialty pharmacies) Median Weighted by Medicaid Volume for all Pharmacies (including specialty pharmacies) Mean Weighted by Medicaid Volume for all Pharmacies (including specialty pharmacies) Value for Average Dispensing Fee Estimate of Fixed Component for Dispensing Fee $8.98 $6.81 $10.12 $7.93 $9.47 $7.29 $11.57 $9.35 7

8 Chapter 2: Cost of Dispensing Methodology and Findings The Texas Health and Human Services Commission (HHSC) engaged Myers and Stauffer LC to perform a study of costs incurred by pharmacies participating in the Texas Medicaid pharmacy program to dispense prescription medications. There are two primary components related to the provision of prescription medications: dispensing cost and drug ingredient cost. Dispensing cost consists of the overhead and labor costs incurred by a pharmacy to fill prescription medications. In its final rule to implement provisions of the Deficit Reduction Act of 2005 (DRA), the Centers for Medicare and Medicaid Services (CMS) has provided some basic guidelines for appropriate costs to be reimbursed via a Medicaid pharmacy dispensing fee. CMS guidelines state: Dispensing fee means the fee which (1) Is incurred at the point of sale or service and pays for costs in excess of the ingredient cost of a covered outpatient drug each time a covered outpatient drug is dispensed; (2) Includes only pharmacy costs associated with ensuring that possession of the appropriate covered outpatient drug is transferred to a Medicaid recipient. Pharmacy costs include, but are not limited to, reasonable costs associated with a pharmacist s time in checking the computer for information about an individual s coverage, performing drug utilization review and preferred drug list review activities, measurement or mixing of the covered outpatient drug, filling the container, beneficiary counseling, physically providing the completed prescription to the Medicaid beneficiary, delivery, special packaging, and overhead associated with maintaining the facility and equipment necessary to operate the pharmacy; and (3) Does not include administrative costs incurred by the State in the operation of the covered outpatient drug benefit including systems costs for interfacing with pharmacies. 1 Further guidance from CMS regarding pharmacy dispensing came in proposed rule published on February 2, In this proposed rule, CMS did not 1 See Medicaid Program; Prescription Drugs; Final Rule. Federal Register, 72: 136 (17 July 2007), p. 39,240. These guidelines are codified at 42 CFR Note that the proposed rule included within Medicaid Program: Covered Outpatient Drugs Federal Register, 77:22 (2 Feb. 2012), p. 5361, would modify this definition, but only by modifying the term dispensing fee to be professional dispensing fee stating that Professional dispensing fee means the professional fee which Otherwise, the modified definition for a dispensing fee would remain the same. 8

9 fundamentally alter the definition for the components of the dispensing fee as currently codified at 42 CFR other than to replace the terminology of dispensing fee with professional dispensing fee. However, the importance of the pharmacy dispensing fee was highlighted in this proposed rule in conjunction with a proposed change in the basis for ingredient reimbursement from the currently defined estimated acquisition cost (EAC) 3 to actual acquisition cost (AAC). 4 The requirement that state Medicaid agencies should more closely match their ingredient reimbursement to actual acquisition cost highlights the importance of the professional dispensing fee. CMS states in the proposed rule: we feel that this change from dispensing fee to professional dispensing fee reinforces our position that once the reimbursement for the drug is properly determined, the dispensing fee should reflect the pharmacist s professional services and costs associated with ensuring that possession of the appropriate covered outpatient drug is transferred to a Medicaid beneficiary. Therefore, as States change their payment for ingredient cost, we also propose to require States to reconsider the dispensing fee methodology consistent with the revised requirements. (p. 5326) Methodology of the Dispensing Cost Survey In order to determine costs incurred to dispense pharmaceuticals to members of the Texas Medicaid pharmacy program, Myers and Stauffer utilized a survey method consistent with CMS guidelines for the components of a pharmacy dispensing fee in 42 CFR and the methodology of previous surveys conducted by Myers and Stauffer in several states. Survey Distribution Myers and Stauffer obtained from the HHSC a list of pharmacy providers currently enrolled in the Texas Medicaid pharmacy program. According to the provider list, there were 4,433 pharmacy providers enrolled in the program that were located within the state of Texas. Myers and Stauffer used a random sampling approach to develop a sample of 2,000 pharmacy providers which were to be included in the pharmacy cost of dispensing survey. This same sample of 2,000 pharmacies were also included in a survey of drug purchase invoices as 2 See Medicaid Program; Covered Outpatient Drugs. Federal Register, 77: 22 (2 February 2012) p See 42 CFR for definition of the EAC ( the agency's best estimate of the price generally and currently paid by providers for a drug marketed or sold by a particular manufacturer or labeler in the package size of drug most frequently purchased by providers ) and 42 CFR for upper limits of payment that incorporate the current EAC requirement. 4 In the proposed rule, AAC is defined as the agency s determination of the pharmacy providers actual prices paid to acquire drug products marketed or sold by specific manufacturers. (p. 5359). 9

10 part of a separate project performed by Myers and Stauffer for the HHSC to evaluate the ingredient portion of pharmacy reimbursement. Cost of dispensing survey materials were mailed to all 2,000 sampled pharmacy providers on October 25, Each surveyed pharmacy received a copy of the cost survey (Exhibit 1), instructions for the survey (Exhibit 2) and letters of explanation from the HHSC (Exhibit 3a and Exhibit 3b). Concerted efforts to encourage participation were made to enhance the survey response rate. A survey help desk was provided by Myers and Stauffer. A toll-free telephone number and address were listed on the survey form and pharmacists were instructed to call or to resolve any questions they had concerning completion of the survey form. The survey instructions and a letter of explanation offered pharmacy owners the option of having Myers and Stauffer complete certain sections of the survey for those that were willing to submit copies of financial statements and/or tax returns. For convenience in completing the cost of dispensing survey, the survey forms were also made available in electronic formats and pharmacies were provided with options to report data using templates based on Excel spreadsheets. Reminder letters were sent on November 25, 2013 to surveyed pharmacies (Exhibit 4a and Exhibit 4b). Additional letters were sent on December 9, 2013 with a further reminder and an extension of the original due date of December 6, 2013 to January 8, 2013 (Exhibit 5a and Exhibit 5b). Providers were given instructions to report themselves as ineligible for the survey if they met certain criteria. Pharmacies could be deemed ineligible if they had closed their pharmacy, had a change of ownership or had less than six months of cost data available (e.g., due to a pharmacy that recently opened, or changed ownership). Of the 2,000 surveyed pharmacies, 34 pharmacies reported themselves to be ineligible to participate. Surveys were accepted through the end of January As indicated in Table 2.1, 1,263 surveyed pharmacies submitted a usable cost survey for this study resulting in a response rate of 64.2%. Some of the submitted cost surveys contained errors or did not include complete information necessary for full evaluation. For cost surveys with such errors or omissions, the pharmacy was contacted for clarification. There were limited instances in which issues on the cost survey were not resolved in time for inclusion in the final analysis. There were 41 incomplete surveys received on or before January 31, 2014 that were eventually determined to be unusable because they were substantially incomplete or missing essential information. These issues could not be resolved in a timely manner with the submitting pharmacy. These 41 incomplete surveys are not included in the count of 1,263 usable surveys received. 10

11 The following table, 2.1, summarizes the cost of dispensing survey response rate. Table 2.1 Dispensing Cost Survey Response Rate Pharmacy Category Medicaid Enrolled Pharmacies Located in Texas Sampled Pharmacies Receiving Cost Surveys Pharmacies Exempt or Ineligible from Filing Eligible Pharmacies Usable Cost Surveys Received Response Rate Chain 2,921 1, , % Non-chain 1, % TOTAL 4,433 2, ,966 1, % In-State Urban 3,639 1, ,593 1, % In-State Rural % TOTAL 4,433 2, ,966 1, % Tests for Reporting Bias For the pharmacy traits of affiliation (i.e., chain or independent) and location (i.e., urban or rural), the submitted surveys were tested to determine if they were representative of the population of Medicaid provider pharmacies. Since the response rate of the surveyed pharmacies was less than 100 percent, the possibility of bias in the response rate should be considered. To measure the likelihood of this possible bias, chi-square ( 2 ) tests were performed. A 2 test evaluates differences between proportions for two or more groups in a data set. Of the 1,263 usable cost surveys, 935 were from chain pharmacies and 328 were from non-chain pharmacies. There was a response rate of 70.0% for chain pharmacies compared to a response rate of 52.0% for independent pharmacies. The results of the 2 test indicated that the difference in response rates was statistically significant. A 2 test was also performed with respect to the urban versus rural location of the pharmacy for responding pharmacies that were located in the state of Texas. The results of this test indicated that the difference in response rate for pharmacies in urban versus rural locations (a response rate of 63.3% for urban pharmacies compared to a response rate of 68.4% for rural pharmacies) was not statistically significant. Review Procedures All submitted surveys were subjected to review procedures to ensure that data used to determine the cost of dispensing at each pharmacy was complete, accurate and reasonable. 11

12 Desk Reviews A desk review was performed for 100% of surveys received. This review identified incomplete cost surveys and pharmacies submitting these incomplete cost surveys were contacted by telephone and/or to obtain information necessary for completion. The desk review process also incorporated a number of tests to determine the reasonableness of the reported data. In many instances, pharmacies were contacted to correct or provide confirmation of reported survey data that was indicated for review as a result of these tests for reasonableness. On-Site Field Examinations In addition to the desk review procedures, 32 pharmacies were selected for an on-site field examination. The selection was primarily random, but geographic location was also taken into consideration. A letter was sent to each selected pharmacy notifying them of selection for an on-site visit, the time period during which the field examination would take place, and the necessary data to have available (Exhibit 6). Each pharmacy was then contacted by telephone for further explanation of the field examination and scheduling of a specific time and date. An examination file was prepared for each of the pharmacies containing a uniform field examination program, a copy of the completed reviewed cost survey and other necessary work papers. On-site field examinations were performed by accounting staff who were familiar with the pharmacy cost of dispensing survey process. Review steps that were performed included validating the prescription volume and expenses reported on the submitted survey. Additional procedures were used to verify the statistics used to allocate expenses (i.e., area ratio, sales ratio and time worked in the prescription department). Following the actual visit to the pharmacy, work papers were completed and reviewed for quality assurance. Although the on-site field examinations resulted in adjustments that had both a positive and negative impact on the cost of dispensing calculated for the individual pharmacies that were visited, the overall conclusion of the on-site field examination process was that there was no significant bias in overstating or understating costs reported on the cost survey (Exhibit 7). Cost Finding Procedures For all pharmacies, the basic formula used to determine the average dispensing cost per prescription was to calculate the total cost related to dispensing prescriptions and divide that amount by the total number of prescriptions dispensed: 12

13 Total (Allowable) Dispensing Related Cost Average Dispensing Cost Total Number of Prescriptions Dispensed Determining the average dispensing cost potentially becomes complex since not all reported costs are strictly related to the prescription dispensing function of the pharmacy. Most pharmacies are also engaged in lines of business other than the dispensing of prescription drugs. For example, many pharmacies have a retail business with sales of over-the-counter (OTC) drugs and other non-medical items. Some pharmacies are involved in the sale of durable medical equipment. The existence of these other lines of business necessitates that procedures be used to isolate the costs involved in the prescription dispensing function of the pharmacy. Cost finding is the process of recasting cost data using rules or formulas in order to accomplish an objective. For this study, the objective is to estimate the cost of dispensing prescriptions to Medicaid members. To accomplish this objective, some pharmacy costs must be allocated between the prescription dispensing function and other business activities. This process identified the reasonable and allowable costs necessary for prescription dispensing to Medicaid members. Dispensing cost consists of two main components: overhead and labor. The cost finding rules employed to determine each of these components are described in the following sections. Overhead Costs Overhead cost per prescription was calculated by summing the allocated overhead of each pharmacy and dividing this sum by the number of prescriptions dispensed. Overhead expenses that were reported for the entire pharmacy were allocated to the prescription department based on one of the following methods: All, or 100% overhead costs that are entirely related to prescription functions. None, or 0% overhead costs that are entirely related to non-prescription functions. Area ratio prescription department floor space (in square feet) divided by total floor space. Sales ratio prescription sales divided by total sales. Overhead costs that were considered entirely prescription-related include: Prescription department licenses. Prescription delivery expense. 13

14 Prescription computer expense. Prescription containers and labels (For many pharmacies the costs associated with prescription containers and labels are captured in their cost of goods sold. Subsequently, it was often the case that a pharmacy was unable to report expenses for prescription containers and labels. In order to maintain consistency, a minimum allowance for prescription containers and labels was determined to use for pharmacies that did not report an expense amount for containers and labels. The allowance was set at the 95 th percentile of prescription containers and labels expense per prescription for pharmacies that did report prescription containers and labels expense: $0.50 per prescription). Certain other expenses that were separately identified on Lines 22a-22r of Page 4 5 of the cost survey (Exhibit 1). Overhead costs that were not allocated as a prescription expense include: Income taxes 6 Bad debts 7 Advertising 8 Charitable Contributions 9 5 Other expenses were analyzed to determine the appropriate basis for allocation of each expense: sales ratio, area ratio, 100% related to dispensing cost or 0% (not allocated). 6 Income taxes are not considered an operational cost because they are based upon the profit of the pharmacy operation. Although a separate line was provided for the state income taxes of corporate filers, these costs were not included in this study as a prescription cost. This provides equal treatment to each pharmacy, regardless of the type of ownership. 7 The exclusion of bad debts from the calculation of dispensing costs is consistent with Medicare cost reporting principles. See Provider Reimbursement Manual, CMS Pub.15-1, Section 304. The allowance of unrecovered costs attributable to such bad debts in the calculation of reimbursement by the Program results from the expressed intent of Congress that the costs of services covered by the Program will not be borne by individuals not covered, and the costs of services not covered by the Program will not be borne by the Program. It is recognized that some bad debts may be the result of Medicaid co-payments that were not collected. However, it was not possible to isolate the amount of bad debts attributable to uncollected Medicaid co-payments from the survey data. Additionally, there may be programmatic policy reasons to exclude uncollected Medicaid co-payments from the calculation of the cost of dispensing. Inclusion of cost for uncollected co-payments in the dispensing fee might serve to remove incentives for pharmacies to collect Medicaid co-payments when applicable. Given that co-payments were established to bring about some measure of cost containment, it may not be in the best interest of a Medicaid pharmacy program to allow uncollected co-payments to essentially be recaptured in a pharmacy dispensing fee. 8 The exclusion of most types of advertising expense is consistent with Medicare cost reporting principles. See Provider Reimbursement Manual, CMS Pub. 15.1, Section Costs of advertising to the general public which seeks to increase patient utilization of the provider's facilities are not allowable. 9 Individual proprietors and partners are not allowed to deduct charitable contributions as a business expense for federal income tax purposes. Any contributions made by their business are deducted along with personal contributions as itemized deductions. However, corporations are allowed to deduct contributions as a business expense for federal income tax purposes. Thus, while Line 14 on the cost report recorded the business contributions of a corporation, none of these costs were allocated as a prescription expense. This provides equal treatment for each type of ownership. 14

15 Certain costs reported on Lines 22a through 22r of Page 5 of the cost survey (Exhibit 1) were occasionally excluded if the expense was not related to the dispensing of prescription drugs. In the following Table 2.2, measurements are provided for certain expenses that were not included in the cost of dispensing. For all of the expenses below, average cost per prescription was calculated using a sales ratio as the basis for allocation. Table 2.2 Non-Allocated Expenses per Prescription Mean Weighted by Medicaid Expense Category Volume Bad Debts $0.491 Charitable Contributions $0.008 Advertising $0.165 n= 1,263 pharmacies The remaining expenses were assumed to be related to both prescription and nonprescription sales. Joint cost allocation is necessary to avoid understating or overstating the cost of filling a prescription. Overhead costs allocated on the area ratio include: Depreciation Real estate taxes Rent 10 Repairs Utilities The costs in these categories were considered a function of floor space. 11 The floor space ratio was increased by a factor of 2.0 from that reported on the 10 The survey instrument included these special instructions for reporting rent: Overhead costs reported on the cost report must be resulting from arms-length transactions between non-related parties. Related parties include, but are not limited to, those related by family, by business or financial association, and by common ownership or control. The most common nonarms-length transaction involves rental of property between related parties. The only allowable expense of such transactions for cost determination purposes would be the actual costs of ownership (depreciation, taxes, interest, etc., for the store area only). This treatment of related-party expenses is consistent with Medicare cost reporting principles. See Provider Reimbursement Manual, CMS Pub. 15-2, Section 3614: Cost applicable to home office costs, services, facilities, and supplies furnished to you by organizations related to you by common ownership or control are includable in your allowable cost at the cost to the related organizations. However, such cost must not exceed the amount a prudent and cost conscious buyer pays for comparable services, facilities, or supplies that are purchased elsewhere. 11 Allocation of certain expenses using a ratio based on square footage is consistent with Medicare cost reporting principles. See Provider Reimbursement Manual, CMS Pub. 15-2, Section

16 original cost survey to allow for waiting and counseling areas for patients and for a prescription department office area. The resulting ratio was adjusted downward, when necessary, not to exceed the sales ratio (in order to avoid allocating 100% of these costs in the instance where the prescription department occupies the majority of the area of the store). Overhead costs allocated using the sales ratio include: Personal property taxes Other taxes Insurance Interest Accounting and legal fees Telephone and supplies Dues and publications Labor Costs Labor costs are calculated by allocating total salaries, payroll taxes, and benefits based on the percent of time spent in the prescription department. The allocations for each labor category were summed and then divided by the number of prescriptions dispensed to calculate labor cost per prescription. There are various classifications of salaries and wages requested on the cost survey (Lines 1a to 4 of Page 6 of the cost survey Exhibit 1) due to the different cost treatment given to each labor classification. Although some employee pharmacists spent a portion of their time performing nonprescription duties, it was assumed in this study that their economic productivity when performing nonprescription functions was less than their productivity when performing prescription duties. The total salaries, payroll taxes, and benefits of employee pharmacists (Lines 2a to 2j of Page 6 of the cost survey Exhibit 1) were multiplied by a factor based upon the percent of prescription time. Therefore, a higher percentage of salaries, payroll taxes, and benefits was allocated to prescription labor costs than would have been allocated if a simple percent of time allocation were utilized. Specifically, the percent of prescription time indicated was adjusted by the following formula: Example: An employee pharmacist spends 90 percent of his/her time in the prescription department. The 90 percent factor would be modified to 95 percent: (2)(0.9)/(1+0.9) = Thus, 95 percent of the reported salaries, payroll taxes, and benefits would be allocated to the prescription department. It should be noted that most employee pharmacists spent 100 percent of their time in the prescription department. 16

17 (2)(%Rx Time) (1 (%Rx Time)) The allocation of salaries, payroll taxes, and benefits for all other prescription employees (Lines 3a to 3h of Page 6 of the cost survey Exhibit 1) was based directly upon the percentage of time spent in the prescription department as indicated on the individual cost survey. For example, if the reported percentage of prescription time was 75 percent and total salaries were $10,000, then the allocated prescription cost would be $7,500. Limits on Reasonable Cost As is often the case in cost-finding procedures used for setting reimbursement rates for Medicaid services, Myers and Stauffer applied reasonable limits to certain reported expenses. The procedures that were applied were intended to ensure that the calculated cost of dispensing was not misstated in the case of related party transactions (e.g., owner s compensation), unrealistic labor allocations or expenses not considered to be directly related to the cost of dispensing medications to Medicaid members. Owner Compensation Issues The allocation of salaries, payroll taxes, and benefits of the owner pharmacists (Lines 1a to 1e of Page 6 of the cost survey Exhibit 1) was based upon the same modified percentage as that used for employee pharmacists. However, limitations were placed upon the allocated salaries, payroll taxes, and benefits of owner pharmacists. Since compensation reported for owner pharmacists are not expenses that have arisen from arm's length negotiations, they are not similar to other costs. A pharmacy owner has a different approach toward other expenses than toward his/her own salary. Owners may pay themselves above the market costs of securing the services of an employee pharmacist. This excess effectively represents a withdrawal of business profits, not a cost of dispensing. Conversely, some owners may underpay themselves for business reasons, which would also misrepresent the true dispensing cost. A factor considered in determining the allocation of owner's salaries was the variability in productivity. For example, one owner pharmacist may dispense 30,000 prescriptions per year while another may dispense 5,000. Those owner pharmacists who dispensed a greater number of prescriptions were allowed a higher salary than were owner pharmacists who dispensed a smaller number of prescriptions. Since variance is not nearly as great with respect to employee pharmacists, the owner pharmacist's salary was subjected to limits based upon employee pharmacists' salaries per prescription. 17

18 To estimate the cost that would have been incurred had an employee been hired to perform the prescription-related functions actually performed by the owner, a statistical regression technique was used. A bivariate plot shows the correlation between an independent (predictor) variable and a dependent (predicted) variable (Exhibit 8). The upper and lower limits on owner pharmacist salary were determined from a bivariate regression. 13 In order to accurately reflect the trend of decreasing marginal costs with increasing volume, a regression technique that fit the bivariate data to a logarithmic curve was used. The resulting regression equation to predict pharmacist labor cost at varying amounts of work performed is: Labor cost = 46,125 X ln (number of prescriptions dispensed) 323,391 (where ln represents the natural logarithm function) This equation was used to establish limits for allocating owner pharmacist costs. There was variation in actual employee salaries both above and below this regression line. This variation is measured by the equation s standard error of the estimate, $31,617. The standard error of the estimate was used to construct upper and lower limits of owner pharmacist labor cost: Upper Limit = 46,125 X ln (number of prescriptions dispensed) 271,387 Lower Limit = 46,125 X ln (number of prescriptions dispensed) 331,401 These two constraints effectively set upper and lower thresholds at approximately the 40 th and 95 th percentiles of volume adjusted employee salaries. These thresholds allow for a conservative adjustment to owner salaries. An additional constraint is a $209,836 maximum salary and a $63,199 minimum salary. These amounts are set at the 40 th and 95 th percentile of volume adjusted employee salaries. There is no reason to believe that managerial or clerical duties performed by the non-pharmacist owners were more valuable to the prescription dispensing function than for other functions. As with other owners, the amount shown for salaries, payroll taxes, and benefits was not a result of arm's length negotiations. Therefore, an upper limit of $83,200 and a lower limit of $31,200 were placed upon these labor costs. These limits were based on an analysis of salaries of employee pharmacists and were adjusted based on the reported time worked by the owner non-pharmacist. A sensitivity analysis of the owner labor limits was performed in order to determine the impact of the limits on the overall analysis of pharmacy dispensing 13 Employee pharmacist salary per prescription was used to set limitations on owner pharmacist salary estimates due to the arm s length nature and lack of variance in employee productivity compared with owner productivity. The number of prescriptions filled by the owner pharmacist was determined by multiplying the percent of owner-filled prescriptions (Lines 1a- 1e of Page 6 of the cost survey - Exhibit 1) by the total number of prescriptions dispensed (Line a of Page 1 of the cost survey - Exhibit 1). 18

19 cost. Of the 1,263 pharmacies in the cost analysis, owner limits impacted 245 pharmacies, or 19.4%. Of these, 66 pharmacies had costs reduced as a result of application of these limits (on the basis that a portion of owner salary cost appeared to represent a withdrawal of profits from the business), and 179 pharmacies had costs increased as a result of the limits (on the basis that owner salaries appeared to be below their market value). In total, the final estimate of average pharmacy dispensing cost per prescription was decreased by approximately $0.02 as a result of the owner salary limits. 14 Overall Labor Cost Constraints An overall constraint was placed on the proportion of total reported labor that could be allocated as prescription labor. The constraint assumes that a functional relationship exists between the proportion of allocated prescription labor to total labor and the proportion of prescription sales to total sales. It is also assumed that a higher input of labor costs is necessary to generate prescription sales than nonprescription sales, within limits. The parameters of the applied labor constraint are based upon an examination of data submitted by all pharmacies. These parameters are set in such a way that any resulting adjustment affects only those pharmacies with a percentage of prescription labor deemed unreasonable. For instance, the constraint would come into play for an operation that reported 75 percent pharmacy sales and 100 percent pharmacy labor (obviously, some labor must be devoted to generating the 25 percent nonprescription sales). To determine the maximum percentage of total labor allowed, the following calculation was made: 0.3(Sales Ratio) 0.1 ( 0.2 )(Sales Ratio) A sensitivity analysis of the labor cost restraint was performed in order to determine the impact of the limit on the overall analysis of pharmacy cost. The analysis indicates that of the 1,263 pharmacies included in the dispensing cost analysis, this limit was applied to 62 pharmacies. In total, the final estimate of average pharmacy dispensing cost per prescription was decreased by approximately $0.01 as a result of the labor cost restraint. Inflation Factors All allocated costs for overhead and labor were totaled and multiplied by an inflation factor. Inflation factors are intended to reflect cost changes from the 14 The net impact of owner compensation limits excludes one outlier pharmacy. 19

20 middle of the reporting period of a particular pharmacy to a common fiscal period ending December 31, 2013 (specifically from the midpoint of the pharmacy s fiscal year to June 30, 2013 which is the midpoint of the fiscal period ending December 31, 2013). The midpoint and terminal month indices used were taken from the Employment Cost Index, (all civilian, all workers; seasonally adjusted) published by the Bureau of Labor Statistics (BLS) (Exhibit 9). The use of inflation factors allows pharmacy cost data from various fiscal years to be compared uniformly. The majority of submitted cost surveys were based on a fiscal year which ended December 31, Dispensing Cost Analysis and Findings The dispensing costs for surveyed pharmacies are summarized in the following tables and paragraphs. Findings for pharmacies are presented collectively, and additionally are presented for subsets of the surveyed population based on pharmacy characteristics. There are several statistical measurements that may be used to express the central tendency of a distribution, the most common of which are the mean and the median. Findings are presented in the forms of means and medians, both weighted and unweighted. The measures of central tendency used in this report include the following: Unweighted mean: the arithmetic average cost for all pharmacies. Weighted mean: the average cost of all prescriptions dispensed by surveyed pharmacies, weighted by prescription volume. The resulting number is the average cost for all prescriptions, rather than the average for all pharmacies as in the unweighted mean. This implies that low volume pharmacies have a smaller impact on the weighted average than high volume pharmacies. This approach, in effect, sums all costs from surveyed pharmacies and divides that sum by the total of all prescriptions from surveyed pharmacies. The weighting factor can be either total prescription volume or Medicaid prescription volume. Median: the value that divides a set of observations (such as dispensing cost) in half. In the case of this survey, the median is the dispensing cost such that the cost of one half of the pharmacies in the set are less than or equal to the median and the dispensing costs of the other half are greater than or equal to the median. Weighted Median: this is determined by finding the pharmacy observation that encompasses the middle value prescription. The implication is that one half of the prescriptions were dispensed at a cost of the weighted median or less, and one half were dispensed at the cost of the weighted median or more. Suppose, for example, that there were 1,000,000 Medicaid 20

21 prescriptions dispensed by the surveyed pharmacies. If the pharmacies were arrayed in order of dispensing cost, the median weighted by Medicaid volume, is the dispensing cost of the pharmacy that dispensed the middle, or 500,000 th prescription. For both weighted means and weighted medians, the use of Medicaid prescription volume as the weighting factor is particularly meaningful for consideration in determining appropriate reimbursement since it emphasizes the cost of dispensing from those pharmacies that dispense more significant volumes of Medicaid prescriptions. As is typically the case with dispensing cost surveys, statistical outliers are a common occurrence. These outlier pharmacies have dispensing costs that are not typical of the majority of pharmacies. Medians are sometimes preferred to averages (i.e., the arithmetic mean) in situations where the magnitude of outlier values results in an average that does not represent what is thought of as average or normal in the common sense. For all pharmacies, findings are presented in Table 2.3. Table 2.3 Dispensing Cost per Prescription All Pharmacies Dispensing Cost Mean Weighted by Medicaid Volume $11.57 Median Weighted by Medicaid Volume $9.47 n=1,263 pharmacies (Dispensing costs have been inflated to the common point of June 30, 2013) See Exhibit 10 for a histogram of the dispensing cost for all pharmacies. There was a large range between the highest and the lowest dispensing cost observed. However, the majority of pharmacies (approximately 70%) had average dispensing costs between $8 and $16. Exhibit 11 includes a statistical summary with a wide variety of measures of pharmacy dispensing cost with breakdowns for many pharmacy attributes potentially of interest Previously, it was noted that the difference in the survey response rate between chain and independent pharmacies was statistically significant. There was a response rate of 70.0% for chain pharmacies compared to a response rate of 52.0% for independent pharmacies. The difference in the unweighted mean dispensing cost for non-specialty chain and independent pharmacies, $12.10 and $14.94 respectively, was found to be statistically significant (at the 5% level of significance). For purposes of this report, no adjustment to the reported means and medians of dispensing cost has been made as a result of potential response rate bias. 21

22 Specialty Pharmacies Several pharmacies included in the cost analysis were identified as specialty pharmacies, which for purposes of this report are those pharmacies that reported sales for intravenous, home infusion, blood factor and/or other specialty services of 10% or more of total prescription sales. The analysis revealed significantly higher cost of dispensing associated with pharmacies that provided significant levels of these services. 15 A comparison of average dispensing costs for specialty pharmacies and pharmacies that did not offer these specialty services is summarized in Table 2.4. Table 2.4 Dispensing Cost per Prescription - Specialty Versus Other Pharmacies Type of Pharmacy Specialty Pharmacies Other Pharmacies Number of Pharmacies 57 1,206 Average Total Annual Prescription Volume (mean and median) Median Weighted by Medicaid Volume Mean: 78,631 Median: 17,689 $15.35 Mean: 82,465 Median: 71,158 $8.98 n=1,263 pharmacies (Dispensing costs have been inflated to the common point of June 30, 2013) Pharmacies that dispense specialty prescriptions as a significant part of their business often have dispensing costs in excess of those found in a traditional pharmacy. The analyses summarized in Tables 2.5 through 2.8 below exclude the specialty pharmacy providers. In making this exclusion, no representation is made that the cost structure of those pharmacies is not important to understand. However, it is reasonable to address issues relevant to those pharmacies 15 In every pharmacy dispensing study where information on intravenous solution and home infusion dispensing activity has been collected by Myers and Stauffer, such activity has been found to be associated with higher dispensing costs. Discussions with pharmacists providing these services indicate that the activities and costs involved in these specialty prescriptions are significantly different from the costs incurred by the traditional retail or institutional pharmacy. The reasons for this difference include: Costs of special equipment for mixing and storage of specialty products. Higher direct labor costs because most specialty prescriptions must be prepared in the pharmacy, whereas the manual activities to fill traditional prescription are mainly limited to counting pills (or vials, etc.) and printing and affixing the label. There is often inconsistency in the manner in which prescriptions are counted in specialty pharmacies. A specialty pharmacy may mix and deliver many dispensings of a daily intravenous, home infusion or blood factor product from a single prescription, counting it in their records as only one prescription. This results in dispensing costs being spread over a number of prescriptions that is smaller than if the pharmacy had counted each refill as an additional prescription. This latter factor, in particular, can have a dramatic impact on increasing a pharmacy s calculated cost per prescription. 22