Appropriate Use of IMS Information Financial Community Presentation November 12, 2009

Transcription

1 Appropriate Use of IMS Information Financial Community Presentation November 12, 2009

2 Meeting Objective Helping the Financial Community to understand how to optimize use of IMS offerings Share processes and methodologies for the IMS national audits that will aid in better understanding how to interpret published estimates Familiarize you with a few best practice approaches that will enhance your ability to utilize IMS data effectively 2 Confidential and Proprietary to IMS, November 2009

3 Meeting Topics Comparison of National Databases Applying Tools and Resources Applying Best Practices for Cross-Audit Comparisons Impact of Marketplace Pharmacy Practices Other Offerings to support Appropriate Use of IMS data 3 Confidential and Proprietary to IMS, November 2009

4 Comparison of IMS National Databases Key Audit Distinctions

5 Healthcare Map Flow of Sales and Prescriptions through the Healthcare Market DIRECT SALES INDIRECT SALES R X RETAIL/MAIL PHARMACIES R RX/APLD MANUFACTURER WAREHOUSE LONG TERM CARE PHARMACIES R RX DIRECT SALES Supply (DDD /NATIONAL SALES PERSPECTIVES ) NON RETAIL OUTLETS Demand (Xponent /NATIONAL PRESCRIPTION AUDIT ) 5 Confidential and Proprietary to IMS, November 2009

6 The Prescription Database The IMS US prescription database is a representative sample of Retail (Chain, Foodstore, Independent pharmacies), Mail Service and Long Term Care pharmacies. Measures both national and sub national dispensed prescription activity into the hands of the consumer/patient. Data collected from 38,000 sample pharmacies. ~260 million prescriptions collected each month More than 1.1 million prescribers representing 170 unique specialties Information collected from the Prescription Database creates Xponent and NPA. 6 Confidential and Proprietary to IMS, November 2009

7 The Sales Database Measures both national and sub national sales dollars, units and prices of pharmaceutical drugs purchased by Retail and Non-Retail pharmacies. Data collected from both manufacturers direct sales (~ 100 pharmaceutical companies) and indirect sales from over 500 distribution centers. 1.5 billion transactions are processed each year. Information collected from the Sales Database creates DDD and NSP. 7 Confidential and Proprietary to IMS, November 2009

8 Comparison of National Databases Audit Distinctions Characteristic NSP NPA Audit Definition National sales estimates of product packages into retail drugstores, hospitals, clinics, and other nonretail type outlets National estimates of prescriptions, or the rate at which drugs move out of the pharmacy and into the hands of a consumer via formal dispensed prescriptions Distribution Purchases (Supply) Dispensed Rx s (Demand) Source of the Data Channel Coverage Warehouse ship-to invoice data and manufacturer reported direct sales Retail Mail (includes VA mail) LTC (includes VA Nursing Home) Clinics Federal Facilities HMO Home Healthcare Miscellaneous Non-Federal Hospital Pharmacy terminal data Retail Mail (excludes VA mail) LTC (excludes VA Nursing Home) Projections All channels projected, except Mail All channels projected Data Availability Monthly Weekly and Monthly Specialty No Physician Specialty Data Physician Specialty available Reporting Timing reporting Calendar reporting 8 Confidential and Proprietary to IMS, November 2009

9 National Sales Perspectives Calendar for Confidential and Proprietary to IMS, November 2009

10 Appropriate Use of IMS Information Applying Tools and Resources

11 Available resources may explain variances Method IMS Confidence Intervals NPA Monthly Expected Change Tables Recommended Usage Confidence Intervals are built around the current estimate (i.e. +/- from the current estimate) and are good to use if the estimate needs to be compared to another number, such as a target, to evaluate if two values are statistically significantly different. -Reflect the variations created by the number of days for each day of the week, for a given month, weighted by the relative importance of each day of the week. -Major holidays are given their own unique weights. Provide insight into expected month to month (year over year) trends. Historic Variability - month to month/week to week variability IMS Product News Communications If attempting to validate a month-to-month (or week-toweek) change, historic variability will provide a range of potential values May explain data anomalies or market events that are reflected in the data Note: Typically one would not use all methods. Follow the general guidelines above on which tool is appropriate for each situation. 11 Confidential and Proprietary to IMS, November 2009

12 Confidence Intervals 12 Confidential and Proprietary to IMS, November 2009

13 Interpreting data Confidence Intervals What do they tell us? A Confidence Interval provides a range of values that are a statistical tie A value outside a Confidence Interval is unlikely to occur under typical conditions It is said to be statistically significantly different from typically occurring values Provides point in time views of a data point Utilizing Confidence Intervals when using statistical estimates builds robustness into decisions IMS provides confidence interval tables for the national offerings 13 Confidential and Proprietary to IMS, November 2009

14 Understanding estimates based on a sample Building tolerance around estimates X True value (unknown) Statistical Estimate (known) Confidence interval around estimate to account for sample to sample variation: ( X ) 14 Confidential and Proprietary to IMS, November 2009

15 Possible Outcomes 1. If each estimate has a CI about it, then the difference is statistically significant if the two intervals DO NOT OVERLAP ( ) (( )) Ŷ 1 Ŷ 2 2. If each estimate has a CI about it, and one or both estimates fall inside interval of other then the difference is not statistically significant ( (( ) )) Ŷ 1 Ŷ 2 3. If each estimate has a CI about it, and intervals overlap but neither estimate falls inside the interval of the other then the result is inconclusive ( (( ) )) Ŷ 1 Ŷ 2 15 Confidential and Proprietary to IMS, November 2009

16 NPA Confidence Intervals How to apply NPA Confidence Intervals 1. Identify the appropriate Confidence Interval Table a. Identify if the product in question is Branded or Generic, etc. b. Use product volume for data point in question within a particular channel and locate the appropriate row (always default to lower volume) c. Locate Channel in question (note that attached example pulls retail channel table) d. Locate percentage for 2 standard deviations (95% confidence interval) 2. Calculate the upper and lower value and apply to target data point a. Multiply 95% confidence interval to data point in question b. Compare upper and lower bounds to target data point 16 Confidential and Proprietary to IMS, November 2009

17 Confidence Intervals Example Are the October 2008 and November 2008 TRx volumes statistically significantly different? 17 Confidential and Proprietary to IMS, November 2009

18 18 Confidential and Proprietary to IMS, November 2009

19 = 617, % = 617, % 647, ,489 = 631, % = 631, % 19 Confidential and Proprietary to IMS, November 2009

20 20 Confidential and Proprietary to IMS, November 2009

21 CONCLUSION: After applying the upper and lower bounds, it can be concluded that the November 2008 TRx volume is a statistical tie to the prior month, October 2008, and therefore NOT statistically significantly different. Nov. 617,054 Oct. 631,022 Ŷ 2 600, , , ,555 Ŷ 1 21 Confidential and Proprietary to IMS, November 2009

22 NPA Expected Change Tables 22 Confidential and Proprietary to IMS, November 2009

23 Interpreting Data NPA Expected Change Tables Removes the expected month-to-month changes to gauge a true percent change Updated to reflect the new calendar year NPA estimates take into consideration: number of days in each month differences in prescription volume processed through Retail, Mail and Long-Term Care pharmacies according to each day of the week NPA Expected Change can only be used for monthly data, not applicable for weekly trends. 23 Confidential and Proprietary to IMS, November 2009

24 Sample NPA Expected Change Table 24 Confidential and Proprietary to IMS, November 2009

25 NPA Expected Change Tables How to apply NPA Expected Change 1. Identify appropriate Expected Change table Retail, Mail Service or Long-Term Care 2. Calculate month-over-month percent change (Current month Previous month)/previous month 3. Compare month-over-month observed percent change to month-over-month expected change 25 Confidential and Proprietary to IMS, November 2009

26 Expected Change Example What is the true fluctuation in NPA Retail volume from January 2008 to February 2008 if you back out the variance due to the number of days in the month and differences in volume processed by each day of the week? 26 Confidential and Proprietary to IMS, November 2009

27 CONCLUSION: After applying the expected change to the observed growth, the smoothed change for Product A is -4.29%. Instructions 1.Calculate % change over preceding Month. 2.Identify observed % change and subtract the expected % change from the chart. Example: [(-9.39) (-5.10) = -4.29%] 3. Result is a smoothed % change which eliminates the effect of changing the distribution of weekdays. 27 Confidential and Proprietary to IMS, November 2009

28 Historic Variability 28 Confidential and Proprietary to IMS, November 2009

29 Interpreting data What is a historic variability trend? Provides a range of potential values when attempting to validate month-to-month or week-to-week changes. Historic variability creates tolerance bands that account for the natural variation over time. 29 Confidential and Proprietary to IMS, November 2009

30 Interpreting the data Normal variations exist in trends over time Quantify historic month-to-month % variation; and use to distinguish if current period change is reflective of normal trend Month to Month Percent Change in TRx Volume: Aug 06 to Jul 08 (Product A Oral Solid) 15.0% 10.0% % Change in TRxs 5.0% 0.0% -5.0% -10.0% -15.0% Aug-06 Sep-06 Oct-06 Nov -06 Dec-06 Jan-07 Feb-07 Mar-07 Apr-07 May -07 Jun-07 Jul-07 Aug-07 Sep-07 Oct-07 Nov -07 Dec-07 Jan-08 Feb-08 Mar-08 Apr-08 May -08 Jun-08 Jul Confidential and Proprietary to IMS, November 2009

31 Apply Resources Creating month-to-month historic variability in Excel Column A: Sort months in ascending order Column B: TRx Volume Column C: Calculate % change (see formula) Use Average and Standard Deviation (two Excel built-in functions in functions list) and create your upper and lower bounds for acceptable variability in the month-to-month change. All the computations are listed here. 31 Confidential and Proprietary to IMS, November 2009

32 Historic Variability Example Is the published IMS TRx growth estimate for the September 2008 data month consistent with the historic variability for this product? 32 Confidential and Proprietary to IMS, November 2009

33 Apply Resources How to calculate and apply Historic Variability 1. Calculate month over month or week over week percentage change: (Current time period Previous time period)/previous time period 2. Calculate average percentage change based on 24 historical time periods, not including the period in question [AVERAGE formula] 3. Calculate standard deviation [STDEV formula] 4. Calculate upper and lower variability tolerance based on 2 standard deviations [Average +/- (2*stdev)] 5. Compare upper and lower variability tolerance to percent change 33 Confidential and Proprietary to IMS, November 2009

34 CONCLUSION: The variance is within acceptable historic variability for the period (i.e., -8.14% is between -9.2% and 15.7%). 34 Confidential and Proprietary to IMS, November 2009

35 Interpreting data Examining Historic Variability Use for any estimate tracked over time: volumes, shares, dollars, quantities (extended units) Use for tracking time periods: monthly, weekly, daily (examine by day of week), Use for tracking products: Generics, Brands, or USC Therapeutic Categories, etc. Remove obvious differences among time periods prior to creating bounds Use with at least 24 time periods and two (2) Standard Deviations Our recommendations for tighter bounds: Trend % change as opposed to volume when tracking new products (get closer to normal approximation) Apply expected change tables when tracking monthly Rx volumes of drugs in NPA (accounts for shorter months, etc.) For monthly NSP dollars and extended units, divide by number of weeks in month and apply bounds to average weekly numbers (to account for calendar month) 35 Confidential and Proprietary to IMS, November 2009

36 IMS Communications 36 Confidential and Proprietary to IMS, November 2009

37 Review IMS Communications If, after applying the resource applications, questions still exist, review IMS Product News: Market event Enhancement Advisory Follow-up If applicable, apply guidance, and analyze impact reporting results 37 Confidential and Proprietary to IMS, November 2009

38 Interpreting IMS National Estimates Step 1: Data is published Step 2: Apply Resources Step 3: IMS Communication Step 4: Contact IMS Receive published estimate Interpret data in light of market & company knowledge Calculate Confidence Intervals Apply NPA Expected Change Tables, if applicable Analyze data in light of Historic Variability If questions exist: Review published Product News communications Apply Guidance and analyze impact reporting results, if applicable If questions exist: Contact your IMS Account representative for initiation of a data investigation 38 Confidential and Proprietary to IMS, November 2009

39 Applying Tools and Resources Closing summary A true picture of the market is extremely intricate and multifaceted due to multiple sources of variation IMS is well positioned to handle the market intricacies and variation IMS captures a robust and broad sample Utilizes our patented estimation method Employs data collection and communication best practices In conjunction with IMS, the Financial Community can employ practices that: Distinguish normal variation from true performance indicators by using tolerance bands Set expectations that volume fluctuations, regardless of root cause, that fall within normal expected bands do not represent statistical differences 39 Confidential and Proprietary to IMS, November 2009

40 Guidelines for National Database Comparisons Applying Best Practices

41 Guidelines for Comparing Audits Market Definitions When comparing sales and prescription data, always ensure that market definitions, channel definitions, product/product groupings, forms, or strengths match across IMS databases. 41 Confidential and Proprietary to IMS, November 2009

42 Guidelines for Comparing Audits Metrics Dollars are not the recommended measurement to use when comparing sales and prescription data. Example Price per Extended Unit: Product A 10mg Tablet Price Amount Source Pharmacy Retail $2.20 NPA Price Pharmacy $1.65 NSP Retail (Transaction) Acquisition Price Wholesaler Acquisition Cost (WAC) $1.75 DDD 42 Confidential and Proprietary to IMS, November 2009

43 Guidelines for Comparing Audits Metrics, continued Prescriptions are NOT the recommended measure to use when comparing sales and prescription data Extended Units ARE the recommended measure when comparing sales and prescription activity Prescription Extended Units generally correlate to NSP Extended Units Extended Units may vary depending upon pharmacy reporting practices 43 Confidential and Proprietary to IMS, November 2009

44 Example of Data Measure Distinctions NSP and NPA: Impact of Extended Unit (EU) definitions Pack configuration: 12 tubes of.25 grams NSP Extended Units= # of grams NPA Extended Units= # of tubes Once NPA is factored to represent grams, NSP and NPA show similar volume trends Source: NSP Extended Units (EU RET) and NPA Extended Unit Total Dispensed Prescriptions (EUTRx) 44 Confidential and Proprietary to IMS, November 2009

45 Guidelines for Comparing Audits Time Periods A three-month rolling average is the recommended time period to use Formula: (3 months summed) divided by 3 = three-month rolling average (January + February + March) / 3 (February + March + April) / 3 (March + April + May) / 3 45 Confidential and Proprietary to IMS, November 2009

46 Guidelines for Comparing Audits Channels Traditional Retail Channels are suggested for cross-audit comparisons Traditional Retail Channels (Chain, Independent, Foodstore pharmacies) are reported across both the sales and prescription databases and are by definition, similar. Long-Term Care Channels are reported across both the sales and prescription databases and are by definition, similar. Mail Service definitions across databases are dissimilar. Projections VA Mail 46 Confidential and Proprietary to IMS, November 2009

47 Case Study Comparing data from two audits: Dollars vs. Dispensed Rxs Why are NPA and NSP volumes and trends NOT SIMILAR? Source: NSP Dollars (DOL TOT) and NPA Total Dispensed Prescriptions (TRx) 47 Confidential and Proprietary to IMS, November 2009

48 Applying Guidelines Step 1: Step 2: Step 3: Step 4: Compare Metrics Time Periods Channels Identify key audit distinctions Market Definitions Product/Product Groupings Form and Strength Use Extended Units Ensure Extended Unit definitions across databases are the same Recommend three month rolling average Compare Channels Retail: Chain Independent Foodstores LTC Mail dissimilar Projections VA Mail differences 48 Confidential and Proprietary to IMS, November 2009

49 Applying Guidelines Step 1 Step 1: Step 2: Ste 3: Step 4: Compare Metrics Time Periods Channels/ Subcats Identify key audit distinctions Market Definitions Product/Product Groupings Form and Strength NSP Dollars NPA Prescriptions Schedule Calendar Month Retail, Mail, LTC & Non- Retail VA Mail Included Unprojected Mail Retail, Mail, LTC VA Mail Excluded Projected Mail 49 Confidential and Proprietary to IMS, November 2009

50 Applying Guidelines Steps 2 and 3 Step 2: Metrics Step 3: Time Periods Use Extended Units Ensure Extended Unit definitions across databases are the same Recommend threemonth rolling average 50 Confidential and Proprietary to IMS, November 2009

51 Case Study Comparing data from two audits NSP reporting is vs. NPA calendar reporting. NSP Dollars NPA TRXs Source: NSP Dollars (DOL TOT) and NPA Total Dispensed Prescriptions (TRx) 51 Confidential and Proprietary to IMS, November 2009

52 Apply Appropriate Use Metrics and Time Periods Extended Units and Average Rolling Months applied Source: NSP Extended Units (EU TOT) and NPA Extended Unit Total Dispensed Prescriptions (EUTRx) 52 Confidential and Proprietary to IMS, November 2009

53 Applying Guidelines Step 4 Step 1: Step 2: Step 3: Step 4: Compare Metrics Time Periods Channels Compare Channels Retail: Chain Independent Foodstores LTC Mail dissimilar Projections VA Mail differences 53 Confidential and Proprietary to IMS, November 2009

54 Apply Appropriate Use Channels Like Channels applied (Chain, Independent, Foodstore, Mail & LTC) Source: NSP Extended Units (EU TOT) and NPA Extended Unit Total Dispensed Prescriptions (EUTRx) 54 Confidential and Proprietary to IMS, November 2009

55 Further Break-out of Like Channels Retail Channel view (Chain, Independent and Foodstore) Source: NSP Extended Units (EU TOT) and NPA Extended Unit Total Dispensed Prescriptions (EUTRx) 55 Confidential and Proprietary to IMS, November 2009

56 Applying Guidelines Step 1: Step 2: Step 3: Step 4: Compare Metrics Time Periods Channels Identify key audit distinctions Market Definitions Product/Product Groupings Form and Strength Use Extended Units Ensure Extended Unit definitions across databases are the same Recommend three month rolling average Compare Channels Retail: Chain Independent Foodstores LTC Mail dissimilar Projections VA Mail differences 56 Confidential and Proprietary to IMS, November 2009

57 NPA Rx Measures Marketplace Dynamics of Pharmacy Practices

58 IMS definition of New, Refill and Total Rx NRx Prescriptions issued by the pharmacy with a unique prescription number A new prescription may or may not represent new therapy for the patient receiving the prescription Refill Prescriptions issued under the same prescription number as an original (new) prescription previously recorded by the reporting pharmacy. Refill prescriptions are necessary for seeing the total prescription volume for a product or market TRx New prescriptions plus refill prescriptions reported by pharmacies. 58 Confidential and Proprietary to IMS, November 2009

59 Standard pharmacy New Rx designation practice Prescriber writes a new Rx Patient presents new prescription at pharmacy with refills Pharmacist Inputs required adjudication and tracking information Pharmacy System Generates a unique Rx number and flags that as New Rx Most Pharmacies track all subsequent Refill Rx identified with the same prescription number 59 Confidential and Proprietary to IMS, November 2009

60 However some exceptions do exist Some pharmacy business processes may cause variation in NRx measures that may not be indicative of a new prescription being presented at the pharmacy Examples 1. Pharmacy assigns a new prescription number to a refill when dispensed NDC number changes. This may occur when: Pharmacy changes the source of a generic or brand product (repackager) resulting in a new dispensed NDC Pharmacy stocks a different package size than traditionally stocked, resulting in a different NDC Patient may switch from brand to generic, or vice-versa Managed Care guidelines may require a New Rx yearly, or twice per year 2. Retail, Mail or LTC Supplier changes dispensing location, resulting in an assignment of a new prescription number 60 Confidential and Proprietary to IMS, November 2009

61 Common Drivers of Pharmacy Practices Influences on New to Refill Ratios Drivers of this practice may include: Supply shortage of a specific manufacturer s generic product; in some cases this will result in the pharmacy sourcing the product from another manufacturer New pharmacy product package size introduced into the market A mail service pharmacy reallocates workload by transferring scripts to satellite pharmacies What to anticipate: A temporary increase in NRx measures with no corresponding increase in TRx measures 61 Confidential and Proprietary to IMS, November 2009

62 Recommendations To the degree you are aware of this type of market event, use prior or subsequent NRx/RRx ratios Use the average of the most recent 6-8 non-impacted weeks, prior to any NRx market event occurring Exclude holiday weeks, in case there are unusual distribution during these time periods IMS continues to work with our data suppliers to mitigate the impact of these sorts of unusual data handling practices. IMS longitudinal prescription information will assist in minimizing Retail supplier based changes in NRx. Longitudinal prescription services categorize prescriptions using anonymous patient IDs, measuring New to Brand (NBRx) or New Therapy Start prescriptions. These services offer alternatives that alleviate NRx fluctuations and enable greater insight into drivers of TRx performance 62 Confidential and Proprietary to IMS, November 2009

63 Other IMS Offerings New To Brand Portfolio Profiler NSP and NPA Pricing Measures Database

64 NPA New to Brand Measuring true new prescription activity A national key performance indicator that integrates industry-leading NPA with de-identified patient-level data Provides enhanced visibility to the volume of true, first-time prescriptions on a brand Only longitudinally-derived metric offered across all products and markets Accessible through investigator.web and Dataview Additional, more granular insight provides a lead indicator of a brand s ability to capture true new business 64 Confidential and Proprietary to IMS, November 2009

65 Case Example The full impact of the Avandia safety news played-out in two weeks through the New-to-Brand metric, while the impact as measured by NRx and TRx is still playing-out Avandia Share 12% 9% TRx share loss 2.96% pts NRx share loss 3.8% pts NBRx share loss 5.4% pts 6% 3% The NBRx share loss represents 8,073 NBRx prescriptions per week 0% Jul Aug Sep Sep Oct Nov Dec 1 06 Dec Jan Feb 2 07 Feb March April April May June 8 07 June July Aug Aug Sep Oct Nov 2 07 Nov Dec Jan Total Rx New Rx New to Brand 65 Confidential and Proprietary to IMS, November 2009

66 IMS MIDAS Portfolio Profiler (Launching in December 2009) Tracks Global Company Performance, Market Trends, and enables users to project potential merger impact Global IMS MIDAS Data available at 4 levels & 3 measures Corporation/ Country/ International-Product/ ATC3 Therapeutic Category USD (US Dollars)/ LCD (Local Currency Dollars)/ SU (Standard Units) Breakout coverage of 15 markets Core: EU5 (Top 5 European countries); US, Canada & Japan Pharmerging: BRICK TM (Brazil, Russia, India, China, Korea; Turkey & Mexico) ROW - Rest of World bucket (Quarterly Only) Data updated monthly where available & for all countries quarterly Monthly data points will be updated weekly until all monthly panels are live Hosted and secured on the IMS Customer Portal Delivery Online Graph, Grid and Data Manipulation Export to Excel, PowerPoint and PDF files 66 Confidential and Proprietary to IMS, November 2009

67 Dashboard View 67 Confidential and Proprietary to IMS, November 2009

68 Country View (filtered) 68 Confidential and Proprietary to IMS, November 2009

69 Data Grid View 69 Confidential and Proprietary to IMS, November 2009

70 NPA Pricing Measures Database The National Prescription Audit has key insights into actual Retail/Consumer pricing. Using Retail Costs derived from pharmacy terminals, NPA provides a valuable reference point for comparing Retail Costs per Rx generated by products within a competitive mix. Delivered as a Dataview database via File Express through the IMS Customer Portal Includes all markets Most recent three years of pricing history Available as a Monthly, or Quarterly Dataview database subscription 70 Confidential and Proprietary to IMS, November 2009

71 NPA Pricing Measures - Sample Report 71 Confidential and Proprietary to IMS, November 2009

72 Summary Identified key distinctions across the National Sales Perspectives and the National Prescription Audit databases Introduced and applied new tools & resources Confidence Intervals Expected Change Historic Variability trend Reviewed and applied guidelines for database comparisons Highlighted marketplace dynamics of pharmacy practices Outlined some additional IMS offerings to support Financial Community customers in the use of IMS data 72 Confidential and Proprietary to IMS, November 2009

73 Questions? IMS Client Service Team: Cathy Jefferson Service Manager Tel. (610) Kathy Constable Senior Client Service Analyst Tel. (610) Confidential and Proprietary to IMS, November 2009