The Economic Freedom Network		Prescription Drug Prices in Canada and the United States—Part 3 Retail Price Distribution [Previous] [Contents] [Next] Analysis Savings Table 4 shows the savings that could be obtained by a shopper who moves from the sampled pharmacy in his home area with the highest prices to that with the lowest prices. As well, it shows the maximum and minimum savings from cross-border shopping by an American patient who goes to the adjacent Canadian area. For example, a shopper from the pharmacy with the highest priced Celebrex® in the Washington area, US$92.59, who travels to the British Columbian pharmacy with the lowest priced Celebrex®, US$29.82, would save $62.77. However, the Washington customer who already shops at the pharmacy with the cheapest Celebrex® there, US$75, who travels to the most expensive British Columbian pharmacy, will save only US$36.62. In all cases, there is a large difference in the savings earned by the American customer, depending on the pharmacies from which, and to which, he travels. Table 4: Savings from domestic and cross-border bargain hunting (US dollars, 30-day supply)
	Celebrex® 200mg			Lipitor® 40 mg			Paxil® 20 mg
	Domestic	Max	Min	Domestic	Max	Min	Domestic	Max	Min
British Columbia	$8.56			$15.08			$9.45
Manitoba	$7.52			$6.93			$9.86
Ontario	$7.08			$8.96			$7.45
Washington	$17.59	$62.77	$36.62	$23.59	$76.90	$38.23	$16.71	$54.52	$28.36
North Dakota/Minnesota	$25.10	$63.40	$30.77	$32.00	$76.28	$37.35	$27.55	$60.39	$22.98
New York	$21.31	$67.23	$38.84	$21.10	$73.96	$43.90	$18.62	$56.82	$30.75

In some cases, a consumer can save about as much in absolute terms by bargain hunting at home as he can by crossing the border, depending on where he starts. The Midwestern shopper for Celebrex® saves US$30.77 by going from the cheapest pharmacy in the North Dakota and Minnesota area to the most expensive pharmacy in the Manitoba area; he saves US$25.10 by going from the most expensive pharmacy in the North Dakota and Minnesota area to the cheapest in the same area. The shopper for Paxil® in the same area saves at least US$22.98 by crossing the border but saves up to US$27.55 by bargain-hunting at home. (This does not disguise the fact that consumers at the sampled pharmacies can always increase their savings by crossing the Canadian border.)

Furthermore, the shopper who crosses the border incurs greater costs than the one who shops at home because he probably has to drive farther, wait at the border-crossing, exchange money, and get his prescription co-signed by a Canadian doctor. This means that the actual savings by crossing the border are less than indicated above. For example, the cost of using RxPassport's bus service is US$99, which eliminates more than the savings from crossing from Washington to British Columbia to get one month's supply of any one of these drugs. Patients making their own arrangements would probably face higher costs.

Statistically estimated prices

Although prices were not gathered from every pharmacy in each area, observations from the randomly selected pharmacies can be used to infer prices throughout each area. Table 5 shows certain estimated price parameters for each area. These estimates are made easier because the samples suggest that the prices for each drug in each area can be described by the normal (or bell-shaped) distribution.⁷

Table 5 supports the hypothesis that American prices on these drugs are, on average, higher than Canadian prices. However, it also demonstrates that the price of prescription drugs differs significantly from one area to another within one country. We can say with great confidence that:

• The average retail prices for Celebrex® and Paxil® are less in the North Dakota--Minnesota area than in the Washington and New York areas;
• The average retail price for Lipitor® is less in the North Dakota--Minnesota area than in the New York area;
• The average retail price of all three drugs in the Ontario area is higher than in the two other Canadian areas, although the difference is probably not financially significant for consumers.

Furthermore, using the rule of thumb for normal distributions, we can estimate the distribution of prices across pharmacies within one area. Table 6 shows the estimated number of pharmacies in each area that have prices within given ranges.

Table 5: Estimated population means and standard deviations (US dollars, 95% confidence interval)

Table 6: Estimated number of pharmacies in an area within a given range of prices (rounded to nearest US dollar)

Prices and the characteristics of pharmacies

Because the survey was done by telephone, the researchers had a limited ability to assess the characteristics of individual pharmacies that might give them the ability to charge higher prices. Free delivery was expected to be a service that would be priced into the prescriptions. Regression analysis indicates that this is the case. In the three of the four areas where the regression was significant at the 5% level, free delivery explains about one-fifth of the higher price between pharmacies who offer the service and those who do not. Table 7 summarizes the equations.

Regression equations for the Washington area and the North Dakota--Minnesota area were not statistically significant. To interpret these equations, imagine a patient armed with a prescription for one month's worth of the three drugs in our sample. If she were in the British Columbia area, we would expect the total cost of the prescription to be about US$7 higher if she filled it at a pharmacy that offered free delivery instead of one that did not. In the New York area, we would expect the difference to be about US$15. Of course, there are large variances around these estimates.

On the other hand, this service factor probably explains more of the price variance than can be captured in an equation. In the real world, the option of prescription-delivery is not simply free versus not free. There are a variety of delivery services offered: free only during certain times of day, free only to regular customers, free on an ad-hoc basis depending on how busy the pharmacy is, and so on. Unfortunately, this sophistication cannot be captured in the equation.

Furthermore, ordinary least squares regression equations were estimated of the form: A = k + cB, where A was the price of one of the three drugs in an area and B was the price of one of the other two. Given three drugs and six areas, this resulted in 18 equations. Of the 18, only one was not significant at the 5% level. For the other 17, the coefficients ranged from US$0.51 to US$1.51 and the R²'s from 0.13 to 0.97. For example, if a pharmacy in the Washington area charges US$1 more for a 30-day supply of Paxil® than another does, we would expect that pharmacy also to charge US$0.51 more for a 30-day supply of Celebrex®.

This may support the claim that service characteristics cause different prices in different pharmacies. We would expect to see consistently higher prices for drugs in pharmacies with superior service. However, it could also reflect higher prices in pharmacies in areas with higher incomes.

Table 7: Ordinary least squares regressions of total prescription cost versus
free delivery as dummy variable

Limits of the analysis

Besides the inability to capture sophisticated differences in types and quality of services of individual pharmacies, the study has other limits:

• This study only looks at three drugs from among thousands of prescription products of different dosages sold in the United States and Canada.
• Some pharmacies did not give meaningful answers and were not included in the results. The usual answer given was that the pharmacy's staff was too busy to give prices on the telephone. However, some pharmacies indicated that it was against store policy. This may indicate that the pharmacies that opted out had higher prices than those that participated, resulting in selection bias. This question was not resolved.
• For some jurisdictions, it was against policy to report the dispensing fee separate from the total price. This prevented an analysis of the fee as a factor in prices.

[Previous] [Contents] [Next]

info@fraserinstitute.ca
4th Floor, 1770 Burrard Street, Vancouver, BC, Canada, V6J 3G7
Tel: (604) 688-0221 Fax: (604) 688-8539 Book Orders: 1-800-665-3558 ext. 580

You can contact us at the above email address for any comments or information requests. Please report any dead links or technical problems.