Report 2: Economic Burden of Illness in Canada, 2005–2008 – EBIC drug expenditures, 2005–2008

Report 2: EBIC Drug Expenditures, 2005-2008

1. Background

Drug expenditure estimates comprise public and private costs associated with prescription and non-prescription (i.e. over-the-counter) drugs purchased in retail stores.  Estimates represent the final costs to consumers, including dispensing fees, markups and appropriate taxes.  Drugs dispensed in hospitals and other institutions are excluded; drug expenditures in hospitals are captured under the hospital care expenditures cost component of EBIC (5).

The EBIC drug expenditure estimates include prescription drug costs only; non-prescription drug costs could not be allocated across EBIC categories (diagnostic category/subcategory, sex, age group and province/territory). This report describes the data sources and methods used to derive the 2005-2008 drug expenditure estimates. It also presents and discusses the results and the limitations of the data used.

2. Data Sources

Data were obtained from two IMS Brogan (a division of IMS Health Inc.) datasets: the 2006-2008 Canadian Disease and Therapeutic Index (CDTI) and the 2005-2008 CompuScript (CS).Footnote 54

2.1 Canadian Disease and Therapeutic Index

The CDTI is a survey that provides information on the drug prescribing patterns of 652 office-based physicians across Canada (41).Footnote 55 It collects information on patient demographic characteristics (e.g. sex and age), diagnosis (coded using ICD version 9) and drugs prescribed (e.g. product, strength, form, dosage, new/continued therapy).  The CDTI does not include data for the territories, and data for the Prairies (Manitoba, Saskatchewan, Alberta) and Maritimes (Newfoundland and Labrador, Nova Scotia, Prince Edward Island, New Brunswick) are grouped as regions instead of by individual province (41,42).

The CDTI uses the Universal Classification System (USC) to standardize and categorize all drugs according to product type and therapeutic class. The USC is a five-digit code classifying drugs along four levels of categorization, USC2 being the broadest and USC5 the most specific. Example 1 illustrates the associated USC codes of a drug used for bronchial therapy (41).

Example 1:  USC code for inhaled steroids for bronchial therapy
USC Code Class # Digits Description
28000 USC2 2 Bronchial therapy
28300 USC3 3 Asthma
28310 USC4 4 Asthma therapy
28312 USC5 5 Inhaled steroids

2.2 Compuscript

The CS contains information, for nearly 70% of all pharmacies across Canada, on total prescription drug costs (retail price plus dispensing fees) and total drug prescriptions (volume of prescriptions) sold in retail pharmacies across Canada, excluding the territories (41,42). In the CS dataset, total prescription drug costs and drug prescriptions are captured by USC and province.

3. Method

3.1 EBIC 2006-2008 Drug Expenditure Methods

The fields in the CDTI used to produce EBIC estimates were USC5 code, ICD-9 code, sex, age, region (Maritimes, Quebec, Ontario, Prairies and British Colombia) and drug use.Footnote 56 Similarly, the CS fields used in the analysis were USC code, province, total number of drug prescriptions and total prescription drug costs. The CDTI and the CS were merged, using the USC5 code and region/province fields, to create a CDTI-CS database.Footnote 57 Example 2 illustrates a simplified example for one USC5-province group.

Example 2: CDTI-CS Database Post-Merge
USC5 Code (CS and CDTI) Province/Region
(CS/CDTI)
ICD-9 Code (CDTI) Sex (CDTI) Age (CDTI) Drug Use (CDTI) Total No. of Drug Prescriptions (CS) Total Drug Costs (CS)
Note: The ICD-9 code may not always be the same for all records within a USC5-province group.
28312 Ontario 493 Male 14 50 20 1000
28312 Ontario 493 Female 35 25 20 1000
28312 Ontario 493 Female 20 25 20 1000

As shown in Example 2, every record with the same USC5 and province field will be matched with the same total number of drug prescriptions and total drug costs. After the CDTI-CS database had been created, the total drug prescriptions and total drug costs for each USC5-province group were distributed across the CDTI-CS records using the drug use distribution. Example 3 illustrates the process using the same numerical example as Example 2.

Example 3: CDTI-CS Database After the Distribution of Totals
USC5 Code (CS and CDTI) Province/Region
(CS/CDTI)
ICD-9 Code (CDTI) Sex (CDTI) Age (CDTI) Total No. of Drug Prescriptions (CS) Total Drug Costs (CS)
Note: The ICD-9 code may not always be the same for all records within a USC5-province group.
28312 Ontario 493 Male 14 10 500
28312 Ontario 493 Female 35 5 250
28312 Ontario 493 Female 20 5 250

As the CDTI contains a region field and not a province field, each prairie and maritime province was assumed to have the same ICD code, sex, age and drug use distribution as its associated region. For example, if 10% of drug use in the CDTI prairie data was attributable to males aged 15-34 years for the ICD-9 code 493, Alberta, Saskatchewan and Manitoba would each have 10% of provincial drug costs attributed to males aged 15-34 years for the ICD-9 code 493.

After the total drug prescriptions and total drug costs had been distributed across records in each USC5-province group, the costs were aggregated by EBIC diagnostic category/subcategory, sex, EBIC age group (0-14 years, 15-34 years, 35-54 years, 55-64 years, 65-74 years, 75+ years) and province. EBIC drug expenditures were not estimated for the territories as the CDTI and CS data sources do not hold information on these jurisdictions.

3.2 Redistribution of 2006-2008 Drug Expenditures for Records with Unknown Age and/or Sex

In the CDTI, a small percentage of records were missing values for sex (2.4%) and age (2.0%). Therefore, the costs associated with these records could not be distributed across sex and age categories. Although the number of records with missing data values was small, it was decided to distribute the costs associated with these records across sex and age categories using alternative methods, as described in sections 3.2.1 to 3.2.3. A hypothetical example of asthma in Ontario is used in each section.

3.2.1 Records with missing age values

In the case of records in which a value was present for sex but not for age, costs were redistributed proportionally across all other records with the same sex and a known age, province and diagnostic category using the cost distribution for these records. Example 4 provides a numerical example of the redistribution of costs for records with missing age values. Although EBIC has six age groups, this example assumes only two age groups (35-54 and 55-64 years) for simplification. The cost of $2 million is redistributed to other records with known age and the same sex, diagnosis and province.

Example 4: Cost Redistribution for Data with Missing Age Values
  Total Asthma Expenditures (known by Age and Sex) Asthma Expenditures (Males, Unknown Age) Males, Age 35-54 years Males, Age 55-64 years
Asthma Expenditures % Total Expenditures Asthma Expenditures % Total Expenditures
Cost before redistribution $100M $2M $75M 75% $25M 25%
Cost after redistribution $102M $0M $75M + (75% x $2M) = $76.5M $25M + (25% x $2M) = $25.5M
3.2.2 Records with missing sex values

For records with known age but missing sex values, costs were redistributed to records of both sexes within the same age group, diagnosis and province, using the cost distribution of these records. Example 5 provides a numerical example of the redistribution of costs for records with missing sex values.

Example 5: Cost Redistribution for Data with Missing Sex Values
  Total Asthma Expenditures (known by Age and Sex) Asthma Expenditures (Age 15–34 years, Unknown Sex) Males, Age 15-34 years Females, Age 15-34 years
Asthma Expenditures % Total Expenditures Asthma Expenditures % Total Expenditures
Cost before redistribution $100M $3M $60M 60% $40M 40%
Cost after redistribution $103M $0M $60M + (60% x $3M) = $61.8M $40M + (40% x $3M) = $41.2M
3.2.3 Records with missing age and sex values

For records missing both age and sex values, costs were redistributed across records with known age and sex with the same diagnosis and province, using the cost distribution of these records. Example 6 provides a numerical example of the redistribution of costs for records with missing age and sex values. For simplicity, it was assumed that only two groups of individuals (males aged 55-64 years and females aged 15-34 years) have costs with known sex and age for asthma in the province of Ontario.

Example 6: Cost Redistribution for Data with Missing Age and Sex Values
  Total Asthma Expenditures (known by Age and Sex) Asthma Expenditures (Unknown Age and Sex) Males, Age 55-64 years Females, Age 15-34 years
Asthma Expenditures % Total Expenditures Asthma Expenditures % Total Expenditures
Cost before redistribution $100M $5M $20M 20% $80M 80%
Cost after redistribution $105M $0M $20M + (20% x $5M) = $21M $80M + (80%x $5M)
= $84M

3.3 EBIC 2005 Drug Expenditure Methods

At the time analysis for the current edition of EBIC began, the CS dataset was not available for the year 2005, since IMS Brogan held these data for a period of only 72 months. To obtain total drug costs for 2005 for each USC5-province group, the costing information from the CS 2006-2010 dataset was used. Specifically, 2005 total drug costs by USC5 and province were estimated by multiplying an estimated cost per prescription by the drug prescription totals. The 2005 cost per prescription by USC5 and province was estimated using the average annual growth rate, after adjusting for inflation, of 2006-2010 CS cost per prescription data (total prescription drug costs/total number of prescriptions) for each USC5-province group. The 2005 total drug prescriptions by USC5 and province were also estimated using the average annual growth rate of 2006-2010 CS drug prescription totals for each USC5-province group. Furthermore, the average annual growth rate for total drug costs and total drug prescriptions for each USC5-province group was estimated using at least three years of CS data. Records in the CDTI 2005 with missing values for required fields (ICD-9, sex, age and region) were dropped before this dataset was merged with the CS 2005 (estimated). Once the CS 2005 (estimated) had been merged with the CDTI 2005, 99.5% of CDTI records were matched with a cost; records not matched with a cost were dropped. As in the other years of analysis, the total drug costs and total drug prescriptions for a USC5-province group were then distributed to records within that USC5-province group using the drug use distribution.

4. ResultsFootnote 58

4.1 Expenditures by Diagnostic Category

Table 6 provides an overview of the EBIC 2005-2008 national drug expenditures by diagnostic category.  In 2008, the top five diagnostic categories with the highest expenditures were cardiovascular diseases ($4.3 billion, 15.3%), neuropsychiatric conditions ($3.6 billion, 12.7%), musculoskeletal diseases ($2.0 billion, 7.1%), endocrine disorders ($1.7 billion, 6.2%) and digestive diseases ($1.4 billion, 5.1%).  Together, the costs for these five diagnostic categories represented just over 46% of total drug expenditures.  EBIC unattributable drug expenditures are defined as total NHEX drug expenditures minus total EBIC drug expenditures distributed across categories.  The unattributable amount of EBIC 2008 national drug expenditures was $6.7 billion (24.1%).

4.2 Expenditures by Diagnostic Category and Sex

Table 7 illustrates EBIC 2008 drug expenditures by diagnostic category and sex. In 2008, 45.9% ($9.7 billion) and 54.1% ($11.5 billion) of expenditures were attributable to males and females respectively. The three diagnostic categories with the largest expenditures for males were cardiovascular diseases ($2.4 billion), neuropsychiatric conditions ($1.5 billion) and endocrine disorders (0.9 billion). For females these were neuropsychiatric conditions ($2.0 billion), cardiovascular diseases ($1.9 billion) and musculoskeletal diseases ($1.3 billion).

In 2008, the five diagnostic categories with the largest difference in cost distributions across the sexes were nutritional deficiencies (21.5% male and 78.5% female), other neoplasms (25.7% male and 74.3% female), congenital anomalies (30.1% male and 69.9% female), malignant neoplasms (32.3% male and 67.7% female) and musculoskeletal diseases (33.0% male and 67.0% female). The ‘maternal conditions’ category is not included in this ranking because costs are only attributable to females.  Furthermore, estimation of unattributable drug expenditures by sex was not possible.

4.3 Expenditures by Diagnostic Category and Age Group

Figure 16 illustrates EBIC 2008 drug expenditures for each age group.  Individuals aged 0-14 years incurred the lowest percentage of drug expenditures (4.8%) and individuals aged 35-54 years the highest (30.0%).

Figure 17 illustrates EBIC 2008 drug expenditures by diagnostic category and age group for the five most costly diagnostic categories. Expenditures were highest for individuals aged 35-54 years, except for the cardiovascular diseases category.

5. Limitations

EBIC 2005-2008 drug expenditure estimates reflect only prescription drugs and exclude non-prescription (i.e. over-the-counter) drugs. Therefore, EBIC may underestimate total drug expenditures, as shown by the 24%-25% unattributable percentage of drug expenditures across the years 2005-2008. Information on non-prescription drug expenditures by diagnostic category would have provided value by reducing the unattributable amount of EBIC drug expenditures. Information on non-prescription drugs may also be important since the cost distribution may be considerably different than that of prescription drugs. With non-prescription drug costs distributed, the costs for certain diagnostic categories may increase relative to other diagnostic categories. The collection of data for non-prescription drugs may be difficult to obtain as these drugs are often used to treat multiple health conditions. Additionally, if the costs associated with drugs prescribed in hospitals could be separated from the hospital care cost component this may also affect the distribution of drug costs by EBIC categories (as well as the distribution of hospital care expenditures by EBIC categories).

The greatest limitation of the EBIC 2005-2008 drug expenditure estimates pertains to the use of the CDTI to distribute total drug expenditures across EBIC categories (diagnosis, sex, age and province).  The CDTI was the only data source that linked drug costs to diagnosis for all health conditions. However, the CDTI surveyed only 652 physicians (1% of the physician population) 2 days every quarter (41).Footnote 59 Given the CDTI’s small sample size and reporting period, the cost distribution of EBIC 2005-2008 drug expenditures may not reflect the true burden across EBIC categories. Additionally, the CDTI data were grouped for provinces in the Prairies and Maritimes, when in reality drug use patterns may vary among these provinces within each region. Although the CDTI captures the pattern of drugs that physicians prescribe for patients, there is no information to determine whether the written prescriptions were actually filled.

As the CS database was not available for 2005, the 2006-2010 CS databases were used to estimate total drug prescriptions and total drug costs for each USC5-province group; these estimates may not represent true values.

Drug costs for records in the 2006-2008 CDTI-CS database with missing sex and/or age data were distributed across the corresponding records with known values; unfortunately, misrepresentation of drug costs by EBIC category may have occurred.

The available data sources did not contain information on prescription drug expenditures for the territories. Several methods were considered to estimate these expenditures; however, they were considered inappropriate for the current edition of EBIC.Footnote 60 The primary concerns were related to differences in population, illness and injury distributions and price variations between the territories and the other provinces/regions.

Given the stated limitations, the distribution of 2005-2008 EBIC drug expenditures may not reflect the true cost distribution by EBIC categories. It is expected that drug expenditures for some diagnostic categories/subcategories (perhaps sex and age groups also) were either over or underestimated; the direction and magnitude of these inaccuracies is unknown.

6. Conclusion

EBIC 2005-2008 drug expenditures were estimated by diagnostic category/subcategory, sex, age group and province/territory.  In 2008, 75.9% of total drug expenditures were attributable across EBIC categories.  The three diagnostic categories with the highest expenditures were: cardiovascular diseases ($4.3 billion, 15.3%), neuropsychiatric conditions ($3.6 billion, 12.7%) and musculoskeletal diseases ($2.0 billion, 7.1%).  Males accounted for 45.9% of 2008 drug expenditures while females accounted for 54.1%. EBIC 2008 drug expenditures were lowest and highest for individuals aged 0-14 years (4.8%) and 35-54 years (30.0%) respectively.

The unattributable amount of drug expenditure estimates is influenced by non-prescription drug costs, as EBIC drug expenditure estimates include only the costs associated with prescription drugs (out of hospital). Additionally, drug expenditures for the territories are not included in the EBIC 2005-2008 estimates. EBIC drug expenditure estimates were distributed across diagnostic category/subcategory, sex, age group and province using a survey that captured drugs dispensed by physicians. Unfortunately, this survey had a small sample size and sampling period, and therefore EBIC estimates may misrepresent the true distribution of drug expenditures across EBIC categories.



Report a problem or mistake on this page
Please select all that apply:

Thank you for your help!

You will not receive a reply. For enquiries, contact us.

Date modified: