Understanding Drug Formulary Listing Decisions in Canada: a Logistic Model. Charles Thompson Candidate, MSc. University of Ottawa

Size: px

Start display at page:

Download "Understanding Drug Formulary Listing Decisions in Canada: a Logistic Model. Charles Thompson Candidate, MSc. University of Ottawa"

Owen Alfred Lynch
5 years ago
Views:

1 Understanding Drug Formulary Listing Decisions in Canada: a Logistic Model Charles Thompson Candidate, MSc. University of Ottawa

2 Introduction Expected Results Provincial governments in Canada are responsible for building and maintaining drug formularies lists of pharmaceuticals available to insured citizens at reduced prices to consumers because subsidized by government. Inclusion on a public drug formulary helps shape demand for individual drug products. There is good evidence of significant disagreement between provinces on listing decisions. Implying that access to individual medications differs significantly by jurisdiction. Criteria for provincial formulary listing decisions are not well understood. All provinces save Alberta have declined to discuss listing criteria with researchers (Attaran et al. 2011). 2

3 Literature Review Expected Results Drug policy is a mix of scientific evidence, judgement, altruism, selfinterest and politics superimposed on a complex, semi-rational, overburdened, constantly changing healthcare system. (Laupacis, 2006) In this context it is difficult to say what factors are driving drug reimbursement decisions in Canadian provinces. The big question: which factors most influence the decision of provincial governments to fund or not fund individual medications? 3

4 Introduction Literature Review Anis et al. (2001), Gregoire et al. (2001), MacDonald & Potvin (2004) and Morgan, Hanley, Raymond & Blais (2009) all reported relatively low levels of agreement between provincial formularies. Though Morgan and colleagues note that products listed on all nine (Quebec excepted) provincial formularies accounted for 77% of total retail spending. Gamble et al. (2011) found that agreement with the Common Drug Review ranged between 60.4% and 96.2% with the lowest agreement in Ontario and the highest in Atlantic provinces. Attaran et al. (2011) suggest a similar range of agreement and imply that the rate of provincial agreement with CDR is declining over time. 4

5 Note the wide range of agreement across jurisdictions and time periods. Source: Attaran Cartagena & Taylor, (2011). Note also the pronounced directionality of disagreement towards deficit disagreement especially in the late period. 5

6 1. Lower than expected inter-provincial agreement or agreement with the CDR should not be read to suggest needed medications are being denied to patients based on jurisdiction. 1. Public drug plans are organized differently; a fact that almost certainly has an impact on the decisions that get made. The issue is not that listing decisions are somehow misinformed or illegitimate but rather that the determinants of listing decisions in individual provinces are not well understood. 6

7 Introduction Methodology Logistic regression models were constructed using a purposeful selection of covariates (Hosmer, Lemeslow & Sturdivant, 2013) Dependent Variable: Coverage or Non-Coverage of Product (Extracted from e-formularies). Benefit at any price considered as positive listing outcome. Sample: 206 observations of 178 drug products reviewed by the Common Drug Review between 2003 and 2013 Gives sufficient lead time for provinces to make a listing decision. Based on Gamble et al. (2011) this is a conservative strategy. 7

8 Literature Review Conceptual Development Methodology Variable Name Coding Source CDR Recommendations Binary; 3 variables Conditional, Unconditional and Do Not List CADTH s Common Drug Review Website Clinical Uncertainty Binary CDR Drug Reviews Therapeutic Category Dummy Variables for 10 Categories CDR Drug Reviews Economic Uncertainty Binary CDR Drug Reviews Relative Price Dummy Variables for 5 Categories CDR Drug Reviews Public Drug Expenditures per Capita/GDP in the Year Preceding Coverage Provincial Deficit per Capita/GDP in the Year Preceding Coverage Peer Effects Continuous Continuous Dummy Variables for the Behaviour of 5 Provinces (0 = not covered beforehand; 1 = covered beforehand) CIHI (2014), Statistics Canada Provincial Government Public Accounts, Statistics Canada Provincial Government e-formularies and Personal Communications with Relevant Government Departments 8

9 Due to capacity constraints in Atlantic Canada, certain information could not be retrieved for this study. As a result, peer effects and budgetary factors could not be assessed. 9

10 Introduction Literature Review Conceptual Development Methodology A McFadden s Pseudo R-Squared greater than 0.2 is generally considered to indicate satisfactory model fit. Province McFadden s Pseudo R-Squared BC AB Other regression diagnostics were also assessed. SK MB 0.32 ON NB NS NL PEI

11 BC AB SK MB ON NB NS NL PEI CDR Unconditional CDR Conditional CDR Similar CDR DNL Multiple Indications Price Great Price Mid + Price Low No Price Clinical Uncertainty Economic Uncertainty 11

12 BC AB SK MB ON NB NS NL PEI Analgesia Arthritis - Cardio + Diabetes HIV Infection - CNS + Ophthalmology - Cancer + Rare - Misc

13 BC AB SK MB ON NB NS NL PEI GDP per Capita Drug Expenditure per Capita Surplus/Deficit Peer BC + Peer AB + + Peer SK + + Peer MB Peer ON 13

14 Introduction Literature Review Methodology No smoking gun, but 1. The CDR s influence on listing decision is significant at least on DNL recommendations. Conditional listing recommendations not significant (possibly methodological) Atlantic Canada seems to be especially receptive (note goodness-of-fit). 1. Peer effects tell an interesting story. Ontario not as much as leader as imagined Western provinces influence one another difficult to map 2. At the extremes, relative price seems to have the predictable effect 3. What is going on with Arthritis drugs in Ontario? 1. Budget effects are difficult to decypher inelegant variable selection? 14

15 15

FORMULARIES IN CANADA PART 1: GENERAL OVERVIEW

ALIGNMENT AMONG PUBLIC FORMULARIES IN CANADA PART 1: GENERAL OVERVIEW Published by the Patented Medicine Prices Review Board October 2017 Alignment Among Public Formularies in Canada Part 1: General Overview