Page 7: Human Antimicrobial Use Report – 2014 - Appendix A

Appendix A

Methods for data collection and analysis

Physician diagnosis data

The Canadian Disease and Therapeutic Index (CDTI) is a quarterly profile that provides information about the patterns and treatments of disease encountered by office-based physicians (specialists and general practitioners, including those with offices in hospitals). These data are presented over three geographic regions in Canada: West (British Columbia, Alberta, Saskatchewan, and Manitoba), Central (Québec and Ontario), and East (New Brunswick, Newfoundland and Labrador, Nova Scotia, and Prince Edward Island). The sample of physicians represents all major specialties across Canada.

The data is collected using a two-stage stratified design, first stratifying by region and then by physician specialty. Data from 652 physicians were available in 2014, and projection methods were used to extrapolate data to a universe of approximately 57,218 physicians in Canada. The estimated standard error associated with the extrapolation to overall diagnoses is of 1.6%. Data on physician prescription/recommendations were extracted from the IMS Health Canada Inc. database.

For four consecutive quarters, each physician in the CDTI group maintains a practice diary describing information on every patient visit during a randomly selected 48-hour period within the quarter. Information includes patient age and gender, reason for visit, diagnosis, name(s) of the treatment(s) recommended or discussed (including drugs, referrals, environmental, behavioural, or dietary changes, etc.), desired therapeutic effect(s), and the presence of concomitant therapies. If a patient presents to a doctor's office for multiple "diseases", the practitioner will generate one form for each disease. If a person visits multiple times for the same "disease", it is counted in the system separately every time. CDTI data were used to determine the most common diagnoses, defined by the International Classification of Diseases Ninth Revision System (ICD-9), and associated with antimicrobial drug mentions for sampled physicians.

The total number of antimicrobial recommendations and number of diagnoses per 10,000 inhabitants for a given year, age or region was obtained by dividing by the size of the population in thousands during that year, age group and region (Table A.1 and Table A.2).

Some limitations and caveats should be considered for assessing data from the CDTI dataset:

  • The drugs listed are those that the physician has written or recommended and do not represent actual prescriptions dispensed by pharmacists or products consumed by the patient, as information on patient compliance was not available.
  • Although diagnoses are listed for each visit, diagnoses are not necessarily made at the visit in question, as previous diagnoses are recorded for chronic conditions.
  • Because physicians were instructed to provide one page per diagnosis, it is not possible to determine if other comorbidities influenced the treatment decision.
  • The data does not include patient visits to primary care nurses.
  • Visit counts do not translate into number of patients (multiple visits per individual are possible, as well as multiple diagnoses per individual).
  • Drug recommendations are not necessarily tied to a prescription, as samples may be given by the practitioner. Similarly, patients may choose not to fill a prescription.
  • Physicians also record drugs 'previously ordered and continued' for the diagnosis, which would not necessarily tie to a prescription dispensed.
  • Some drug therapy and diagnosis is under-represented due to self-medication (i.e. over-the-counter products).
  • Reliability of the data is dependent on sampling error, so caution should be taken when interpreting those disease categories with a small sample size.

Data is only available at the regional level, where fluctuations may be more or less obvious and specific information for individual provinces could not be determined.

Community Pharmacy Dispensing Data

Canadian CompuScript (CS) tracks the number and size of prescriptions dispensed by community pharmacies in Canada. Data fields include product name (including manufacturer), form, strength, province, age group, the number of prescriptions dispensed, units of product dispensed, and dollars spent monthly for each year.

The sampling frame (or "universe") for this dataset in 2014 consisted of 5,984 pharmacies, covering approximately 64% all retail pharmacies in the Canadian provinces. IMS Health Canada Inc. uses a method of geospatial projection that creates projection factors for application to all non-participating stores. Projection is performed on the basis of the number of stores in the area, distance between stores, and store size. The projection factor was used to extrapolate the number of prescriptions and units dispensed in the stores actually sampled to that of the "universe" (9,406 pharmacies).

The CS data includes prescriptions dispensed through the Non-Insured Health Benefits program (NIHB) for all provinces. However, these data do not include any prescription information for the Yukon, Northwest Territories, and Nunavut due to their low volumes and distinct prescribing trends. In order to fill this gap, NIHB dispensing data were obtained from the First Nations and Inuit Health Branch (FNIHB) of Health Canada. This program provides coverage for benefit claims for specified range of drugs and other medical related costs for eligible First Nations and Inuit populations. Dispensation information provided by NIHB for the Territories was merged with the CS dataset. Both datasets were assessed together and data were cleaned to ensure consistency of recording, and product comparability between the data sources and products available for sale in Canada (via the Health Canada Drug Product DatabaseFootnoteFootnote e ).Footnote e 

Antimicrobials for parenteral administration have been included from 2010 forward. These data were not available to be added for 2002 – 2009. As such, when comparing data from 2010 – 2014 to previous years, it should be assumed that the data from 2002 – 2009 are a slight under-representation of the total volume dispensed. Therefore, any reductions seen in comparison to data from 2002 - 2009 are expected to be smaller than the actual state, and conversely, increases are expected to be slightly inflated.

Defined daily doses (DDDs) were determined according to the World Health Organization's Collaborating Centre for Drug Statistics Methodology. Temporary DDDs (not yet approved but posted on the World Health Organization website) were used when available. For erythromycin ethylsuccinate, all tablets were classified as erythromycin ethylsuccinate tablets (2 g) and all forms other than tablets were classified as erythromycin (1 g). For oral pediazole, the DDD for sulfonamides, combinations with other antibacterials (excl. trimethoprim) (2 g) was used. For oral administration of penicillin G, the DDD for benzylpenicillin by parenteral route (3.6 g) was used. Where a DDD was not available for both routes of administration, the available DDD was used (e.g., for benzylpenicillin, kanamycin). Care should be taken when interpreting the trends for drugs for which the DDD used for calculations are lower than doses prescribed (e.g. the DDD assigned to doxycycline by the WHO is 0.1g when in practice 0.2g are prescribed). In these cases, it appears as if the antimicrobial was prescribed for longer duration compared to other antimicrobials where the DDD matches the usually prescribed dose.

The total amount of active ingredient was obtained by multiplying the number of extended units (real or corrected) by the strength of the product in grams. For combination drugs, the active ingredients of all antimicrobial components were summed to obtain the total number of active ingredients. However, the amount of active ingredient used in the calculation of the total number of DDDs for combination drugs included only the compounds from which the DDDs were derived. For example, for drugs composed of trimethoprim-sulfamethoxazole, only the total number of grams of sulfamethoxazole was used to compute the number of DDDs.

The total number of DDDs per 1,000 inhabitant-days for a given year was obtained by summing all DDDs for each antimicrobial drug and each year. This number was further divided by the size of the population in thousands during that year provided by either FNIHB or Statistics Canada, divided by the number of days in that year (365 or 366). The total number of prescriptions and total cost per 1,000 inhabitants was obtained by dividing the total number of prescriptions or the total cost by the population size in thousands for each year. The cost was then adjusted for inflation using the Bank of Canada Inflation CalculatorFootnoteFootnote f Footnote f . Population data were obtained from updated and preliminary post-census estimates based on the results of the 2011 Census.

The limitations and caveats that should be considered with the Canadian CompuScript dataset include the following:

  • CompuScript only includes products with Health Canada identification numbers (Drug Identification Number, DIN, some but not all Natural Product Number, NPN and some but not all Product Pin Number, PIN).
  • Product data is tracked by its DIN therefore two products with the same DIN are reported together.
  • Hospital dispensaries are not included.
  • Yukon, Northwest Territories and Nunavut are not included due to their very low volumes and distinct prescribing trends.
  • Changes made to the databases are made to the last 72 months only and outside the six year period it is considered a closed dataset as updates cannot be made.

The limitations and caveats that should be considered with the Territorial NIHB data include the following:

  • The data presented only includes populations covered by the NIHB program. These data therefore do not reflect the entire population of the territories and care should be taken in the interpretation of these data in relation to the provinces.
  • Data on antibiotics dispensed through NIHB nursing stations to NIHB clients are not included.
  • Pharmacies ship products out of their home province into the North. Therefore these claims would be included in the provider's home region and not in the Territorial data.

Hospital Purchasing Data

The Canadian Drugstore and Hospital Purchases Audit (CDH) measures the dollar value and unit volume of pharmaceutical and diagnostic products purchased by nearly all Canadian retail pharmacy outlets and hospitals excluding those in the Yukon, Northwest Territories, and Nunavut. Hospital purchasing information was collected from a representative sample of over 677 hospitals. A method of geographical projection was used to create projection factors for a "universe" of approximately 761 hospitals to reflect all Canadian purchases (including non-participating stores). Data on drug purchases were extracted from the IMS Health Canada Inc. database, which included sales from manufacturers and wholesalers. For the purposes of this report, only data on hospital purchases were included in the analyses as pharmacy information was presented using the CompuScript data.

Only information regarding antimicrobials for systemic use was included in the analysis. Defined daily doses (DDDs) were determined according to the World Health Organization's Collaborating Centre for Drug Statistics Methodology. For erythromycin ethylsuccinate, all tablets were classified as erythromycin ethylsuccinate tablets (2g) and all forms other than tablets were classified as erythromycin (1g). For oral pediazole, the DDD for sulfonamides, combinations with other antibacterials (excl. trimethoprim) (2 g) was used. For oral administration of penicillin G, the DDD for benzylpenicillin by parenteral route (3.6 g) was used. Where a DDD was not available for both routes of administration, the available DDD was used (e.g., for benzylpenicillin, kanamycin). Care should be taken when interpreting the trends for drugs for which the DDD used for calculations are lower than doses used (e.g. the DDD assigned to doxycycline by the WHO is 0.1g when in practice 0.2g are prescribed). In these cases, it appears as if the antimicrobial was used for longer durations compared to other antimicrobials where the DDD matches the usually prescribed dose.

The total amount of active ingredient was obtained by multiplying the number of extended units (real or corrected) by the strength of the product in grams. For combination drugs, the active ingredients of all antimicrobial components were summed to obtain the total number of active ingredients. However, the amount of active ingredient used in the calculation of the total number of DDDs for combination drugs included only the compounds from which the DDDs were derived. For example, for drugs composed of trimethoprim-sulfamethoxazole, only the total number of grams of sulfamethoxazole was used to compute the number of DDDs.

 The total number of DDDs per discharge or 1,000 discharges for a given year was obtained by summing all DDDs for each antimicrobial drug and each year. This number was further divided by the total number of hospital discharges which includes deaths, sign-outs, transfers and in some provinces day surgeriesFootnote g . The total cost associated with antimicrobials purchased by hospitals was adjusted for inflation using the Bank of Canada Inflation Calculator.Footnote h  The use of metronidazole was added in 2005.

The limitations and caveats that should be considered with the CDH dataset include the following:

  • The data is estimated and is not census data.
  • Limited tracking of specific niche markets (due to low volume and/or unique distribution).
  • A small number of products may be excluded due to confidentiality (if they are only sold in one outlet).
  • Some data may be excluded to reflect true market trends (i.e. large stockpiling transactions that occur prior to a potential epidemic).
  • Direct sales for a specific manufacturer may not be available leading to underestimation of a product (higher incidence in the hospital than in drug store purchases).
  • The provinces of Prince Edward Island and Newfoundland and Labrador were grouped due to the small volume within each province.
  • Changes made to the databases are made to the last 72 months only and outside the five year period it is considered a closed dataset as updates cannot be made.
  • Hospital patient days or number of hospital beds was not available for this data; general population information was used for developing rates of purchasing.

All data were analyzed using SAS v. 9.3 (SAS Institute Inc., Cary, NC, USA) and Microsoft Excel 2010 (Microsoft Cor., Redmond, WA, USA).

Classification of antimicrobials

Table A.1: Classification of antimicrobials used in Canada
Therapeutic Group Antimicrobials
Aminoglycosides amikacin
gentamicin
streptomycin
tobramycin
Amphenicols chloramphenicol
Beta-lactamase sensitive & resistant penicillins cloxacillin
flucloxacillin
penicillin g
penicillin v
Carbapenems ertapenem
imipenem and cilastatin
meropenem
First-generation cephalosporins cefadroxil
cefazolin
cephalexin
Fourth-generation cephalosporins cefepime
Glycopeptide antibacterials vancomycin
Imidazole derivatives ketoconazole
metronidazole
Macrocyclic fidaxomicin
Macrolides, lincosamides and streptogramins azithromycin
clarithromycin
clindamycin
erythromycin
lincomycin
spiramycin
Nitrofuran derivatives nitrofurantoin
Other antibacterials bacitracin
daptomycin
fosfomycin
methenamine mandelate
Oxazolidinone linezolid
Penicillins with extended spectrum amoxicillin
amoxicillin and enzyme inhibitor
ampicillin
piperacillin
piperacillin and tazobactam
ticarcillin and clavulanic acid
Polymyxins colistin
Quinolone antibacterials ciprofloxacin
levofloxacin
moxifloxacin
norfloxacin
ofloxacin
Second-generation cephalosporins cefaclor
cefoxitin
cefprozil
cefuroxime
Steroid antibacterial fusidic acid
Streptomycin streptomycin
Sulfonamides sulfadiazine
sulfamethoxazole
sulfamethoxazole and trimethoprim
sulfasalazine
Tetracyclines doxycycline
minocycline
tetracycline
tigecycline
Third-generation cephalosporins cefixime
cefotaxime
ceftazidime
ceftriaxone
Trimethoprim trimethoprim

Population sizes used for analysis of all datasets

Table A.2: Population values for age groups, by province in 2014
Province Population (thousands)
0-14 15-59 60+
British Columbia 677.7 2867.6 1086.0
Alberta 752.5 2690.8 678.4
Saskatchewan 213.0 686.1 226.3
Manitoba 239.1 782.9 260.0
Ontario 2190.3 8540.6 2947.8
Québec 1267.5 5008.2 1939.0
New Brunswick 109.9 451.6 192.4
Nova Scotia 132.8 570.4 239.5
Prince Edward Island   23.2 86.7 36.4
Newfoundland and Labrador 75.9 317.0 134.0
Territories (NIHB-covered) 17.8 42.3 6.0
Canada 5699.8 22044.2 7745.9

References

Page details

Date modified: