The impact of pharmaceutical innovation on the burden of disease in Canada, 2000–2016

Abstract

We perform an econometric assessment of the role that pharmaceutical innovation—the introduction and use of new drugs—has played in reducing the burden of disease in Canada, by investigating whether diseases for which more new drugs were launched had larger subsequent reductions in disease burden. Since utilization of a drug reaches a peak about 12–14 years after it was launched, we allow for considerable lags in the relationship between new drug launches and the burden of disease.

We analyze the impact of new drug launches on a comprehensive measure of disease burden—the age-standardized disability-adjusted life-years lost (DALY) rate—and on its two components: the age-standardized years of life lost (YLL) and years lost to disability (YLD) rates. We also analyze the impact of new drug launches on the number of hospital discharges and on the average length of hospital stays.

The number of DALYs lost is significantly inversely related to the number of drugs that had ever been launched 9–20 years earlier, and the number of YLLs is significantly inversely related to the number of drugs that had ever been launched 11–20 years earlier. The launch of a drug has the largest (most negative) impact on the number of DALYs and YLLs 15 years after it was launched.

The estimates indicate that if no drugs had been launched during 1986–2001, the age-standardized DALY rate would not have declined between 2000 and 2016; it might even have increased. Almost all (93%) of the reduction in DALYs was due to a reduction in YLL. The estimates imply that new drug launches during 1986–2001 reduced DALYs in 2016 by 21%, reduced YLLs in 2016 by 28%, and reduced YLDs in 2016 by 3%.

We estimate that drugs launched during 1986–2001 reduced the number of DALYs lost in 2016 by 2.31 million. Expenditure in 2016 on drugs launched during 1986–2001 per DALY gained in 2016 from those drugs was 2842 USD. Interventions that avert one DALY for less than average per capita income for a given country or region are generally considered to be very cost–effective; Canada's per capita GDP was 42,158 USD in 2016, so our estimates indicate that the new drugs launched during 1986–2001 were very cost–effective, overall.

Moreover, 2842 USD may be an overestimate of the true net cost in 2016 per DALY of drugs launched during 1986–2001. A previous study based on U.S. data showed that about 25% of the cost of new drugs is offset by reduced expenditure on old drugs. Also, our estimates indicate that, if no drugs had been launched during 1986–2001, the average length of 2016 hospital stays would have been about 16% higher. The reduction in hospital expenditure due to shorter average length of stay may have been larger than the expenditure on the drugs responsible for shorter hospital stays.

Highlights

• •We assess the role that the introduction and use of new drugs has played in reducing the burden of disease in Canada.
• •New drug launches during 1986-2001 reduced the age-standardized years of life lost (YLL) rate in 2016 by 28%.
• •Drugs launched during 1986-2001 are estimated to have reduced the number of DALYs lost in 2016 by 2.31 million.
• •Expenditure in 2016 on drugs launched during 1986-2001 per DALY gained in 2016 from those drugs was 2842 USD.
• •If no drugs had been launched during 1986-2001, average length of 2016 hospital stays would have been about 16% higher.

1. Introduction

The health status of Canadians has improved during the 21st century. Life expectancy at birth increased from 79.24 years in 2000 to 82.14 years in 2015. Also, the age-standardized rate of potential years of life lost before age 751 per 100,000 population declined from 4214 during 1999–2003 to 3601 during 2009–2013—a 15% decline (Statistics Canada (2018a)).

Some researchers have argued that biomedical innovation has been the principal cause of recent improvements in health. Fuchs (2010) said that “since World War II … biomedical innovations (new drugs, devices, and procedures) have been the primary source of increases in longevity,” although he did not provide evidence to support this claim. Cutler, Deaton, and Lleras-Muney (2006) performed a survey of a large and diverse literature on the determinants of mortality, and “tentatively identif[ied] the application of scientific advance and technical progress (some of which is induced by income and facilitated by education) as the ultimate determinant of health.” They concluded that “knowledge, science, and technology are the keys to any coherent explanation” of mortality. Other research has shown that most technological progress is “embodied”: to benefit from technological progress, people must use new products and services.2

Most scholars agree with Jones’ (1998, pp. 89–90) statement that “technological progress is driven by research and development (R&D) in the advanced world.” In 1997, the medical substances and devices sector was the most R&D-intensive3 major industrial sector in the U.S.: almost twice as R&D-intensive as the next-highest sector (information and electronics), and three times as R&D-intensive as the average for all major sectors. (National Science Foundation (2017)). According to Dorsey et al. (2010), in 2008, 88% of privately-funded U.S. biomedical research expenditure was funded by pharmaceutical and biotechnology firms; the remaining 11% was funded by medical device firms.

The purpose of this study is to assess econometrically the role that pharmaceutical innovation—the introduction and use of new drugs—has played in reducing the burden of disease in Canada. During the period 1980–2016, drugs with 1404 new ATC codes were launched in Canada: about 39 new ATC codes per year, on average.4 For reasons discussed below, there is likely to be a substantial lag between the launch of a new drug and its maximum impact on the burden of disease, so we will allow for considerable lags in the relationship between new drug launches and the burden of disease.

The analysis will be performed using a difference-in-differences (or two-way fixed effects) research design: we will investigate whether diseases for which more new drugs were launched had larger subsequent reductions in disease burden. This design controls for the effects of general economic and societal factors (e.g. income, education, and behavioural risk factors5), to the extent that those effects are similar across diseases, e.g. smoking increases mortality from respiratory and cardiovascular disease as well as lung cancer, and education reduces mortality from all diseases.

The number of new drug launches varied considerably across diseases.6 Fig. 1 shows the number of chemical substances used to treat 5 diseases that had ever been launched in Canada during the period 1980–2016. These five diseases were selected because an identical number of—six— chemical substances had been launched for each disease by the year 1980. During the next 36 years, 14 new drugs for treating ovary cancer were launched; between 5 and 7 new drugs for treating gonorrhea, bladder cancer, and bipolar disorder were launched; only one new drug for treating gout was launched.

Fig. 1.

Number of (WHO ATC5) chemical substances ever launched, 5 diseases, Canada, 1980–2016.

Source: Author's calculations based on data contained in Health Canada Drug Product Database and Thériaque database.

The primary measure of disease burden we will analyze is the number of disability-adjusted life-years (DALYs) lost, as defined and measured by the World Health Organization (2018a). The DALY is a summary measure that combines time lost through premature death and time lived in states of less than optimal health, loosely referred to as “disability”. The DALY is a generalization of the well-known potential Years of Life Lost measure (YLLs) to include lost good health. One DALY can be thought of as one lost year of ‘healthy’ life, and the measured disease burden is the gap between a population's health status and that of a normative reference population. DALYs for a specific cause are calculated as the sum of the YLLs from that cause and the years of healthy life lost due to disability (YLDs) for people living in states of less than good health resulting from the specific cause. The YLLs for a cause are essentially calculated as the number of cause-specific deaths multiplied by a loss function specifying the years lost for deaths as a function of the age at which death occurs.7

Table 1 shows data on the number of DALYs lost (due to all causes) and population, by age group, in Canada in 2000 and 2016. Almost 9 million DALYs were lost in 2016. As shown in row 8, the crude DALY rate (DALYs lost per 100 population) declined by just 2% between 2000 and 2016. However, the DALY rate generally increases sharply with age (e.g., in 2016, the rate among people age 70 and over was about 3 times as high as the rate among people age 50–59), and the Canadian population is aging: the fraction of the population that was age 60 and over increased from 17% in 2000 to 23% in 2016. The age-standardized DALY rate declined by 14% between 2000 and 2016, and the rates among people age 60 and over declined by 20–22%. We will analyze the impact of new drug launches on the age-standardized DALY rate and on its two components: the age-standardized YLL and YLD rates. We will also analyze the impact of new drug launches on the number of hospital discharges and on the average length of hospital stays.

Table 1.

DALYs lost and population by age group, Canada, 2000 and 2016.

row		2000			2016
	age group (years)	DALYs lost (000s)	Population (000s)	DALYs lost per 100 population	DALYs lost (000s)	Population (000s)	DALYs lost per 100 population	% decline in DALYs lost per 100 population, 2000–2016
1	0–4	227	1793	12.7	213	1929	11.1	13%
2	5–14	194	4096	4.7	161	3867	4.2	12%
3	15–29	744	6248	11.9	798	7038	11.3	5%
4	30–49	1678	9867	17.0	1512	9713	15.6	8%
5	50–59	1079	3609	29.9	1451	5428	26.7	11%
6	60–69	1214	2404	50.5	1688	4285	39.4	22%
7	70+	2574	2718	94.7	3071	4031	76.2	20%
8	Total	7711	30735	25.1	8896	36291	24.5	2%
		% of total			% of total
9	0–4	3%	6%		2%	5%
10	5–14	3%	13%		2%	11%
11	15–29	10%	20%		9%	19%
12	30–49	22%	32%		17%	27%
13	50–59	14%	12%		16%	15%
14	60–69	16%	8%		19%	12%
15	70+	33%	9%		35%	11%
16	Total	100%	100%		100%	100%
17	Age-standardized rate1			25.1			21.5	14%

1. Age-standardized rate based on population age distribution in 2000.

Source: World Health Organization, Disease burden and mortality estimates, http://www.who.int/healthinfo/global_burden_disease/estimates/en/index1.html

In the next section, we will describe the econometric model that we will use to assess the role that pharmaceutical innovation has played in reducing the burden of disease in Canada during the period 2000–2016. The data sources used to estimate this model are discussed on Section III. Empirical results are presented in Section IV. Some implications of the estimates are discussed in Section V. Section VI provides a summary.

2. Methods

To assess the impact that pharmaceutical innovation had on the burden of disease, we will estimate models based on the following 2-way fixed effects equation8:

ln(Y_dt) = β_k ln(CUM_DRUG_d,t-k) + α_d + δ_t + ε_dt (1)

where Y_dt is one of the following variables:

• DALY_dt = the age-standardized rate of DALYs lost due to disease d in year t (t = 2000, 2016)

• YLL_dt = the age-standardized rate of years of life lost due to disease d in year t

• YLD_dt = the age-standardized rate of years of healthy life lost due to disability due to disease d in year t

and.

• CUM_DRUG_d,t-k = ∑_m IND_md LAUNCHED_m,t-k = the number of chemical substances to treat disease d that had been launched in Canada by the end of year t-k (k = 0, 1, 2, …,20)

• IND_md = 1 if chemical substance m is used to treat (indicated for) disease d 9

• IND_md = 0 if chemical substance m is not used to treat (indicated for) disease d

• LAUNCHED_m,t-k = 1 if chemical substance m had been launched in Canada by the end of year t-k

•  = 0 if chemical substance m had not been launched in Canada by the end of year t-k

• α_d = a fixed effect for disease d

• δ_t = a fixed effect for year t

Eq. (1) may be considered a health production function (Koç (2004)), and the number of chemical substances ever launched may be considered a measure of the stock of pharmaceutical “ideas.” Jones (2002) argued that “long-run growth is driven by the discovery of new ideas throughout the world.“10 The log-log specification of eq. (1) incorporates the assumption of diminishing marginal productivity of chemical substance launches: each additional chemical substance launch for a medical condition results in a diminishing absolute reduction in disease burden. If the cost of chemical substance launches for different diseases were equal, it would be socially optimal to equalize the absolute reductions in disease burden across diseases. A small percentage reduction in the burden of a high-burden disease would be as valuable as a large percentage reduction in the burden of a low-burden disease.

Estimates based on eq. (1) will provide evidence about the impact of the launch of drugs for a disease on the burden of that disease, but they will not capture possible spillover effects of the drugs on the burden of other diseases. These spillovers may be either positive or negative. For example, the launch of cardiovascular drugs could reduce mortality from cardiovascular disease, but increase mortality from the “competing risk” of cancer. On the other hand, the launch of drugs for mental disorders could reduce mortality from other medical conditions. Prince et al. (2007) argued that “mental disorders increase risk for communicable and non-communicable diseases, and contribute to unintentional and intentional injury. Conversely, many health conditions increase the risk for mental disorder, and comorbidity complicates help-seeking, diagnosis, and treatment, and influences prognosis.”

Due to data limitations, ln(CUM_DRUG_d,t-k) is the only disease-specific, time-varying regressor in eq. (1). If the data were available, we would like to include other regressors in eq. (1), including (1) disease incidence, and (2) the number of non-pharmaceutical medical innovations (e.g. medical device innovations) that had been launched in Canada. However, there is good reason to believe that failure to control for those variables is unlikely to result in overestimation of the magnitude of β_k; exclusion of those variables may even result in underestimation of the magnitude of β_k. Higher disease incidence is likely to result in both higher disease burden and a larger number of chemical substance launches:

Previous studies have shown that both innovation (the number of drugs developed) and diffusion (the number of drugs launched in a country) depend on market size. Acemoglu & Linn, n.d. found “economically significant and relatively robust effects of market size on innovation.” Danzon, Wang, and Wang (2005) found that “countries with lower expected prices or smaller expected market size experience longer delays in new drug access, controlling for per capita income and other country and firm characteristics” (emphasis added).

Although incidence data are not available for most diseases, annual incidence data for the period 1992–2010 are available for 31 cancer sites (breast, lung, etc.). As expected, there is a significant positive correlation across cancer sites between ln(CASES_st) (where CASES_st = the number of patients diagnosed with cancer at cancer site s in year t) and ln(CUM_DRUG_st) (where CUM_DRUG_st = the number of chemical substances to treat cancer at site s that had ever been launched by the end of year t). But estimates of the equation ln(CUM_DRUG_st) = π ln(CASES_st) + α_s + δ_st + ε_st indicate that the growth rate of CUM_DRUG is uncorrelated across cancer sites with the growth rate of incidence. This suggests that estimates of β_k in eq. (1) are unlikely to be biased by the omission of incidence in that equation.

Failure to control for non-pharmaceutical medical innovation (e.g. innovation in diagnostic imaging, surgical procedures, and medical devices) is also unlikely to bias estimates of the effect of pharmaceutical innovation on the burden of disease, for two reasons. First, as noted earlier, 88% of privately-funded U.S. funding for biomedical research came from pharmaceutical and biotechnology firms (Dorsey et al. (2010)).11 Second, previous research based on U.S. data (Lichtenberg (2014a; 2014b)) indicated that non-pharmaceutical medical innovation is not positively correlated across diseases with pharmaceutical innovation.

The dependent variable of eq. (1) is the log of the level of disease burden in year t. We will use data for two years: 2000 and 2016. Substituting those two values of t into eq. (1) yields:

ln(Y_d,2000) = β_k ln(CUM_DRUG_d,2000-k) + α_d + δ₂₀₀₀ + ε_d,2000 (2)

ln(Y_d,2016) = β_k ln(CUM_DRUG_d,2016-k) + α_d + δ₂₀₁₆ + ε_d,2016 (3)

Subtracting eq. (2) from eq. (3) yields:

Δln(Y_d) = β_k Δln(CUM_DRUG_k_d) + δ’ + ε_d’ (4)

where

• Δln(Y_d) = ln(Y_d,2016/Y_d,2000)

• Δln(CUM_DRUG_k_d) = ln(CUM_DRUG_d,2016-k/CUM_DRUG_d,2000-k)

• δ’ = (δ₂₀₁₆ - δ₂₀₀₀)

• ε_d’ = (ε_d,2016 - ε_d,2000)

Eq. (4) is a simple regression of the 2000–2016 growth in the burden of disease d on the growth in the number of drugs that were used to treat disease d that had ever been launched k years earlier.12 To address the issue of heteroskedasticity,13 eq. (4) will be estimated by weighted least squares, weighting by (Y_d,2000 + Y_d,2016)/2.

The mean (across all medical conditions) 2000–2016 log change in disease burden from drugs launched k years earlier is Δ_k = β_k * mean (Δln(CUM_DRUG_k_d)). The (absolute) reduction in 2016 disease burden from drugs launched between 2000 – k and 2016 – k is (∑_d Y_d,2016) * (1 - (1/exp(Δ_k))).

We will estimate 63 (= 3 * 21) versions of eq. (4): one for each of the three measures of disease burden (DALY, YLL, and YLD) for 21 different lag values (k = 0, 1, …, 20). There is likely to be a substantial lag between the launch of a new drug and its maximum impact on the burden of disease. Utilization of recently-launched drugs tends to be much lower than utilization of drugs launched many years earlier. Evidence about the shape of the drug-age (number of years since launch) drug-utilization profile can be obtained by estimating the following equation:

ln(N_SU_mn) = ρ_m + π_n + ε_mn (5)

where

• N_SU_mn = the number of standard units of molecule m sold in Canada n years after it was first launched (n = 0, 1, …, 20)

• ρ_m = a fixed effect for molecule m

• π_n = a fixed effect for age n

The expression exp(π_n - π₁₄) is a “relative utilization index”: it is the mean ratio of the quantity of a drug sold n years after it was launched to the quantity of the same drug sold 14 years after it was launched. We estimated eq. (5), using annual data for the period 2007–2017 on 721 molecules. Estimates of the “relative utilization index” are shown in Fig. 2. These estimates indicate that utilization of a drug reaches a peak about 12–14 years after it was launched. It is used about twice as much then as it was 4 years after launch14

Fig. 2.

Drug age-utilization profile.

Source: Author's calculations based on data contained in Health Canada Drug Product Database and IQVIA MIDAS database.

Due to gradual diffusion of new drugs, the maximum impact of a drug on disease burden is likely to occur many years after it was launched, but the peak effect could occur either more than or less than 12–14 years after launch. The lag might be longer because some drugs for chronic diseases (e.g. statins) may have to be consumed for several years to achieve full effectiveness. But the lag might be shorter because the impact of a drug on disease burden is likely to depend on its quality (or effectiveness) as well as on its quantity (utilization), and drugs launched more recently are likely to be of higher quality than earlier-vintage drugs. 15^,16

As mentioned earlier, in addition to analyzing the impact of new drug launches on DALYs, we will analyze their impact on the number of hospital discharges and on the average length of hospital stays, by estimating the following equations:

Δln(DISCHARGES_d) = β_k Δln(CUM_DRUG_k_d) + δ’+ ε_d’ (6)

Δln(ALOS_d) = β_k Δln(CUM_DRUG_k_d) + δ’+ ε_d’ (7)

Where.

• Δln(DISCHARGES_d) = ln(DISHARGES_d,2016/DISCHARGES_d,2000)

• Δln(ALOS_d) = ln(ALOS_d,2016/ALOS_d,2000)

• DISCHARGES_dt = the number of hospital discharges for disease d in year t (t = 2000, 2016)

• ALOS_dt = average length (number of days) of hospital stays for disease d in year t

Eqs. (6), (7) will be estimated by weighted least squares, weighting by (DISCHARGES_d,2000 + DISCHARGES_d,2016)/2.

3. Data sources

Disease burden data. Age-standardized rates of DALY, YLL, and YLD, by disease and year, were constructed using data obtained from the World Health Organization's Disease Burden Database (World Health Organization (2018b)).17 The disease classification used in the Disease Burden Database is described in Annex Table A of World Health Organization (2018c). Age-standardized rates (per 100,000 population) of Disability-Adjusted Life-Years lost (DALYs), Years of Life Lost (YLLs), and Years of Healthy Life Lost due to Disability (YLDs), by cause, in 2000 and 2016 are shown in Appendix Table 1.

Drug launch data. Health Canada's Drug Products Database (Health Canada (2018)) was used to determine the years in which (WHO ATC 5th-level) chemical substances first received market authorization in Canada.

Drug indications data. Indications (coded by ICD-10) of chemical substances were obtained from Theriaque, a database produced by the French Centre National Hospitalier d'Information sur le Médicament (2018).18 The number of (WHO ATC5) chemical substances ever launched, by cause, during 1980–2016 are shown in Appendix Table 2.

Table 2.

Estimates of β_k parameters of eq. (4).

A. DALYs (disability-adjusted life-years)
row	lag	Estimate	Std. Err.	t Value	p-value	N	mean (regressor)	βk * mean (regressor)
1	0	−0.287	0.118	−2.44	0.016	133	0.234	−0.067
2	1	−0.007	0.112	−0.06	0.950	131	0.255	−0.002
3	2	−0.071	0.116	−0.61	0.542	130	0.270	−0.019
4	3	−0.160	0.105	−1.52	0.131	130	0.295	−0.047
5	4	−0.255	0.096	−2.66	0.009	129	0.331	−0.084
6	5	−0.176	0.081	−2.19	0.031	128	0.361	−0.064
7	6	−0.094	0.073	−1.28	0.203	128	0.396	−0.037
8	7	−0.128	0.074	−1.74	0.085	127	0.422	−0.054
9	8	−0.137	0.070	−1.95	0.053	125	0.442	−0.061
10	9	−0.257	0.073	−3.51	0.001	124	0.473	−0.122
11	10	−0.246	0.078	−3.16	0.002	123	0.482	−0.118
12	11	−0.251	0.066	−3.82	0.000	123	0.516	−0.129
13	12	−0.247	0.066	−3.74	0.000	122	0.529	−0.131
14	13	−0.307	0.063	−4.88	<.0001	122	0.536	−0.164
15	14	−0.331	0.060	−5.48	<.0001	121	0.551	−0.182
16	15	−0.414	0.072	−5.77	<.0001	119	0.557	−0.231
17	16	−0.334	0.063	−5.27	<.0001	116	0.605	−0.202
18	17	−0.323	0.056	−5.81	<.0001	116	0.637	−0.206
19	18	−0.315	0.055	−5.76	<.0001	116	0.633	−0.200
20	19	−0.311	0.053	−5.83	<.0001	112	0.640	−0.199
21	20	−0.281	0.053	−5.28	<.0001	112	0.649	−0.182
B. YLLs (years of life lost)
row	lag	Estimate	Std. Err.	t Value	p-value	N	mean(regressor)	βk * mean(regressor)
22	0	−0.070	0.140	−0.50	0.619	133	0.285	−0.020
23	1	0.231	0.123	1.88	0.063	131	0.293	0.068
24	2	0.111	0.135	0.83	0.410	130	0.300	0.033
25	3	−0.031	0.125	−0.25	0.804	130	0.331	−0.010
26	4	−0.183	0.114	−1.61	0.111	129	0.371	−0.068
27	5	−0.085	0.092	−0.93	0.357	128	0.409	−0.035
28	6	−0.015	0.088	−0.17	0.864	128	0.446	−0.007
29	7	−0.053	0.090	−0.59	0.555	127	0.477	−0.025
30	8	−0.099	0.087	−1.13	0.262	125	0.495	−0.049
31	9	−0.200	0.092	−2.17	0.033	124	0.533	−0.107
32	10	−0.184	0.096	−1.92	0.058	123	0.536	−0.099
33	11	−0.272	0.082	−3.32	0.001	123	0.568	−0.155
34	12	−0.303	0.085	−3.56	0.001	122	0.576	−0.174
35	13	−0.376	0.082	−4.59	<.0001	122	0.600	−0.225
36	14	−0.419	0.077	−5.41	<.0001	121	0.621	−0.260
37	15	−0.513	0.103	−4.98	<.0001	119	0.635	−0.326
38	16	−0.238	0.087	−2.72	0.008	116	0.722	−0.172
39	17	−0.200	0.078	−2.56	0.012	116	0.788	−0.157
40	18	−0.212	0.077	−2.75	0.007	116	0.781	−0.165
41	19	−0.178	0.074	−2.42	0.018	112	0.806	−0.144
42	20	−0.153	0.070	−2.18	0.032	112	0.807	−0.124
C. YLDs (years lost due to disability)
row	lag	Estimate	Std. Err.	t Value	p-value	N	mean(regressor)	βk * mean(regressor)
43	0	−0.116	0.071	−1.63	0.106	133	0.173	−0.020
44	1	0.014	0.059	0.23	0.815	131	0.210	0.003
45	2	−0.008	0.057	−0.13	0.894	130	0.234	−0.002
46	3	−0.027	0.051	−0.52	0.604	130	0.253	−0.007
47	4	−0.052	0.048	−1.09	0.279	129	0.284	−0.015
48	5	−0.053	0.044	−1.21	0.228	128	0.305	−0.016
49	6	0.008	0.035	0.22	0.827	128	0.337	0.003
50	7	−0.001	0.035	−0.02	0.987	127	0.358	0.000
51	8	0.011	0.026	0.43	0.669	125	0.379	0.004
52	9	−0.064	0.027	−2.37	0.019	124	0.401	−0.026
53	10	−0.058	0.029	−2.00	0.049	123	0.417	−0.024
54	11	−0.045	0.024	−1.92	0.058	123	0.453	−0.020
55	12	−0.042	0.022	−1.90	0.061	122	0.472	−0.020
56	13	−0.045	0.022	−2.01	0.047	122	0.458	−0.021
57	14	−0.042	0.022	−1.90	0.060	121	0.467	−0.020
58	15	−0.068	0.025	−2.73	0.007	119	0.465	−0.032
59	16	−0.068	0.026	−2.65	0.009	116	0.468	−0.032
60	17	−0.072	0.025	−2.85	0.005	116	0.458	−0.033
61	18	−0.063	0.024	−2.61	0.010	116	0.457	−0.029
62	19	−0.066	0.026	−2.56	0.012	112	0.444	−0.029
63	20	−0.056	0.025	−2.21	0.030	112	0.463	−0.026

Drug utilization and expenditure data. Data on the quantity (number of standard units) and value (in USD) of prescription drugs sold in Canada, by chemical substance and year (2007–2017) were obtained from the IQVIA MIDAS database.

Cancer incidence data. Data on the number of new cases of primary cancer, by cancer site and year, were obtained from Statistics Canada (2018c).

Hospitalization data. Data on the number of hospital discharges and average length of stay, by diagnosis, in 2000 and 2016, were obtained from the OECD Health Statistics database (OECD (2018a)). The disease classification scheme is provided in OECD (2018b). The number of hospital discharges and average length of stay (in days), by cause, in 2000 and 2016 are shown in Appendix Table 3.

Table 3.

Estimates of β_k parameters of eqs. (6), (7).

A. Discharges
row	lag	Estimate	Std. Err.	t Value	p-value	N	mean (regressor)	βk * mean (regressor)
1	0	0.236	0.260	0.91	0.369	61	0.199	0.047
2	1	0.266	0.252	1.06	0.294	61	0.211	0.056
3	2	0.256	0.246	1.04	0.301	61	0.223	0.057
4	3	0.164	0.212	0.77	0.443	61	0.254	0.042
5	4	0.149	0.192	0.78	0.439	61	0.292	0.044
6	5	0.101	0.175	0.58	0.566	61	0.316	0.032
7	6	0.092	0.180	0.51	0.612	61	0.327	0.030
8	7	0.109	0.183	0.60	0.552	61	0.355	0.039
9	8	0.141	0.167	0.84	0.403	61	0.380	0.054
10	9	0.116	0.159	0.73	0.470	61	0.422	0.049
11	10	0.117	0.166	0.71	0.483	61	0.424	0.050
12	11	0.092	0.152	0.61	0.547	61	0.466	0.043
13	12	0.086	0.149	0.58	0.564	61	0.491	0.042
14	13	0.096	0.146	0.65	0.516	61	0.512	0.049
15	14	0.115	0.140	0.83	0.411	61	0.539	0.062
16	15	0.167	0.157	1.06	0.294	60	0.555	0.093
17	16	0.151	0.159	0.96	0.343	60	0.565	0.086
18	17	0.095	0.140	0.68	0.499	60	0.602	0.057
19	18	0.053	0.135	0.39	0.698	58	0.608	0.032
20	19	0.075	0.142	0.52	0.602	58	0.606	0.045
21	20	0.080	0.138	0.58	0.567	58	0.619	0.049
B. ALOS
row	lag	Estimate	Std. Err.	t Value	p-value	N	mean(regressor)	βk * mean(regressor)
22	0	−0.431	0.157	−2.74	0.008	61	0.199	−0.086
23	1	−0.417	0.153	−2.73	0.008	61	0.211	−0.088
24	2	−0.452	0.147	−3.08	0.003	61	0.223	−0.101
25	3	−0.418	0.124	−3.36	0.001	61	0.254	−0.106
26	4	−0.383	0.112	−3.41	0.001	61	0.292	−0.112
27	5	−0.341	0.102	−3.33	0.002	61	0.316	−0.108
28	6	−0.316	0.108	−2.94	0.005	61	0.327	−0.103
29	7	−0.317	0.109	−2.89	0.005	61	0.355	−0.112
30	8	−0.302	0.100	−3.02	0.004	61	0.380	−0.115
31	9	−0.267	0.096	−2.79	0.007	61	0.422	−0.113
32	10	−0.268	0.100	−2.67	0.010	61	0.424	−0.114
33	11	−0.242	0.092	−2.63	0.011	61	0.466	−0.113
34	12	−0.237	0.090	−2.63	0.011	61	0.491	−0.116
35	13	−0.265	0.087	−3.03	0.004	61	0.512	−0.136
36	14	−0.271	0.082	−3.29	0.002	61	0.539	−0.146
37	15	−0.311	0.093	−3.35	0.001	60	0.555	−0.173
38	16	−0.315	0.093	−3.38	0.001	60	0.565	−0.178
39	17	−0.278	0.082	−3.39	0.001	60	0.602	−0.167
40	18	−0.282	0.078	−3.62	0.001	58	0.608	−0.171
41	19	−0.275	0.083	−3.30	0.002	58	0.606	−0.166
42	20	−0.251	0.082	−3.06	0.003	58	0.619	−0.155

4. Results

Estimates of β_k parameters of eq. (4) are shown in Table 2. In Panel A of Table 2, the dependent variable is Δln(DALY_d) = ln(DALY_d,2016/DALY_d,2000). The estimate in each row of the table is from a separate model. Rows 1–21 show estimates for assumed values of 0, 1, 2, …, 20 years of the lag (k) from drug launch to disease burden. The point estimates and 95% confidence intervals are also plotted (on an inverted scale) in Panel A of Fig. 3, where solid markers denote significant (p-value < .05) estimates and hollow markers represent insignificant estimates. For k ≤ 8, only 3 of the 9 estimates are statistically significant. This is not surprising since, as discussed earlier, utilization of recently-launched drugs tends to be quite low, and there may also be a lag from drug utilization to disease burden reduction. However, for k ≥ 9, all 12 β_k estimates are negative and statistically significant: the number of DALYs lost is significantly inversely related to the number of drugs that had ever been launched 9–20 years earlier. The magnitude of the estimates tends to increase as k increases, until k = 15, when it starts to decline. The launch of a drug had the largest (most negative) impact on the number of DALYs lost 15 years after it was launched.

Fig. 3.

Estimates of β_k parameters of eq. (4).

Panel A of Fig. 4 shows a comparison of the relative utilization and DALY β_k estimate profiles. Utilization of a drug tends to rise until 12 years after launch, remains stable for 3 years, and then starts to decline. The β_k estimates exhibit some volatility in years 0–6, but then generally increase in magnitude until year 15, and begin to decline the year after utilization begins to decline. The correlation between these two profiles is highly statistically significant (correlation = −0.58; p-value = .006).

Fig. 4.

Comparison of relative utilization and β_k estimate profiles.

Fig. 5 is a bubble plot depicting the relationship across diseases between the 1985–2001 percentage increase in the number of drugs ever launched, and the 2000–2016 percentage change in the age-standardized DALY rate. It is clear from the figure that ischemic heart disease is a highly influential observation. Although the fact that an observation is influential does not necessarily mean that it should be excluded, we estimated the model when that observation was excluded. Exclusion of that observation reduced the point estimate of β₁₅ by 36% (from −0.414 to −0.265), but the estimate remained highly significant (t-value = 4.93; p-value < .0001).19

Fig. 5.

Relationship across diseases between % increase in number of drugs ever launched, 1985–2001, and % change in age-standardized DALY rate, 2000–2016.

Bubble area is proportional to (DALY_c,2000 + DALY_c,2016)/2.

Another apparently influential observation is diabetes. This observation is weakening the relationship between drug launches and DALY reduction: despite a large percentage increase in the number of diabetes drugs, the burden of diabetes did not decline by an unusually large amount. This may be due to a significant increase in the prevalence of diabetes.20 Appendix Fig. 1 shows that the prevalence of diabetes in Canada increased significantly between 2000/2001 and 2008. When we exclude both diabetes and ischemic heart disease from the sample, the point estimate of β₁₅ (−0.385) is very close to the estimate reported in row 16 of Table 2, and is highly significant (t-value = 5.69; p-value < .0001).

We also estimated an alternative functional form (semi-logarithmic as opposed to log-log) of the relationship between drug launches and subsequent DALY reduction. These estimates are shown in Appendix Table 4. Twenty of the 21 coefficients are negative and significant. There is little evidence of an inverted U-shaped profile, and the maximum magnitude of the effect (β_k * mean (regressor)) in the semi-logarithmic model is only 38% as large as the maximum magnitude of the effect in the log-log model. But the fit of the semi-logarithmic model is inferior to the fit of the log-log model. When we include both ln (CUM_DRUG_d,2001/CUM_DRUG_d,1985) and (CUM_DRUG_d,2001 - CUM_DRUG_d,1985) in the model, as suggested by Davidson and MacKinnon (1981), the coefficient on the former regressor is significant, but the coefficient on the latter regressor is not. This indicates that the log-log functional form is more appropriate than the semi-logarithmic functional form.21

Now we will briefly summarize estimates of models of the two components of DALY, YLL and YLD.22 In Panel B of Table 2, the dependent variable is Δln(YLL_d) = ln(YLL_d,2016/YLL_d,2000). Rows 22–42 show estimates for assumed values of 0, 1, 2, …, 20 years of the lag (k) from drug launch to disease burden. The point estimates and 95% confidence intervals are also plotted (on an inverted scale) in Panel B of Fig. 3. For k ≤ 10, only 1 of the 11 estimates is statistically significant. However, for k ≥ 11, all 10 β_k estimates are negative and statistically significant: the number of YLLs is significantly inversely related to the number of drugs that had ever been launched 11–20 years earlier. Once again, the magnitude of the estimates tends to increase as k increases, until k = 15, when it starts to decline. The launch of a drug had the largest (most negative) impact on the number of YLLs lost 15 years after it was launched. Panel B of Fig. 4 shows a comparison of the relative utilization and YLL β_k estimate profiles. Like the DALY β_k estimates, the YLL β_k estimates exhibit some volatility in years 0–6, but then generally increase in magnitude until year 15, and begin to decline the year after utilization begins to decline. The correlation between these two profiles is again highly statistically significant (correlation = −0.78; p-value < .001).

In Panel C of Table 2, the dependent variable is Δln(YLD_d) = ln(YLD_d,2016/YLD_d,2000). The point estimates and 95% confidence intervals are also plotted (on an inverted scale) in Panel C of Fig. 3. For k ≤ 8, none of the 9 estimates are statistically significant. For 9 ≤ k ≤ 14, 3 out of 6 β_k estimates are negative and significant. For 15 ≤ k ≤ 20, 6 out of 6 β_k estimates are negative and significant. This indicates that the number of years lost due to disability was reduced by drugs that had been launched up until 15–20 years earlier.

The estimates in Table 2 imply that most of the DALY reduction from new drug launches was due to a reduction in YLL. The estimates of β₁₅ in rows 16, 37, and 58 imply that new drug launches during 1986–2001 reduced DALYs in 2016 by 21% (= 1 – exp (-0.231)), reduced YLLs in 2016 by 28%, and reduced YLDs in 2016 by 3%.23 There will be additional discussion of the magnitudes of these effects in the next section.

The last estimates we will present are estimates of β_k from the hospital discharges and average length of stay equations, eqs. (6), (7). Panel A of Table 3 shows estimates of β_k from the hospital discharges equation, eq. (6). None of the estimates are statistically significant; we see no evidence that new drug launches reduced the number of people discharged from (or admitted to) hospitals. However, since there is strong evidence that new drug launches reduced mortality, they may have increased the number of people “at risk” of being hospitalized, so new drug launches may have reduced the number of hospital discharges per person at risk of being hospitalized.

Panel B of Table 3 shows estimates of β_k from the average length of hospital stay equation, eq. (7). All 21 estimates are negative and highly significant (p-value ≤ .011), indicating that medical conditions for which there were more new drug launches had smaller increases in average length of stay (ALOS).24 In contrast to the DALY and YLL estimates, the magnitudes of the ALOS estimates are larger for more recent drug launches. Perhaps uptake of new drugs is more rapid among hospitalized patients than it is among other patients. However, the overall impact of new drug launches (β_k * mean(Δln(CUM_DRUG_k))) is highest for k = 16: new drugs launched during 1984–2000 had the largest (most negative) effect on ALOS in 2016. The estimate in row 38 of Table 3 indicates that those drug launches reduced ALOS in 2016 by 16% (= 1 – exp(-0.178)).

5. Discussion

The estimates of the DALY model shown in Panel A of Table 2 and Fig. 3, Fig. 4, in conjunction with other data, may be used to calculate the reduction in DALYs lost in 2016 attributable to previous drug launches and the average cost per DALY gained. DALYs are most significantly inversely related to the number of drugs that had ever been launched 15 years earlier. The estimate of β₁₅ in row 16 of Table 2 indicates that drugs launched during 1986–2001 reduced the mean 2000–2016 log change in DALYs lost by −0.231. This implies that, in the absence of those drug launches, DALYs lost in 2016 would have been 26.0% (= 1 - exp(-0.231)) higher. As shown in Table 1, the total number of DALYs lost in 2016 was 8.896 million, so we estimate that drugs launched during 1986–2001 reduced the number of DALYs lost in 2016 by 2.31 million. If those drugs had not been launched, the total number of DALYs lost in 2016 would have been 11.21 million. Similar calculations based on the YLL and YLD estimates imply that almost all (93%) of the reduction in DALYs was due to a reduction in YLL.

As noted earlier (see Table 1), the age-standardized DALY rate declined by 14.4% between 2000 and 2016. The estimate of the model in row 16 of Table 2 indicates that, if no drugs had been launched during 1986–2001, the age-standardized DALY rate would not have declined; it might even have increased.25 Fig. 6 compares actual % declines in age-standardized DALY, YLL, and YLD rates, 2000–2016 to estimated % declines attributable to drugs launched during 1986–2001.

Fig. 6.

Actual vs. estimated % declines in age-standardized DALY, YLL, and YLD rates, 2000–2016.

According to unpublished IQVIA data, expenditure in Canada in 2016 on drugs launched during 1986–2001 was 6.57 billion USD.26 This expenditure estimate, along with our estimate of the reduction in DALYs lost, implies that pharmaceutical expenditure per DALY gained in 2016 from drugs launched during 1986–2001 was 2842 USD (= 6.57 billion USD/2.31 million DALYs).27 As noted by Bertram et al. (2016), authors writing on behalf of the WHO's Choosing Interventions that are Cost–Effective project (WHO-CHOICE) suggested in 2005 that “interventions that avert one DALY for less than average per capita income for a given country or region are considered very cost–effective; interventions that cost less than three times average per capita income per DALY averted are still considered cost–effective.” Canada's per capita GDP was 42,158 USD in 2016, so these estimates indicate that the new drugs launched during 1986–2001 were very cost–effective, overall.

Several considerations suggest that 2842 USD may be an overestimate of the true net cost in 2016 per DALY of drugs launched during 1986–2001. First, that estimate is based on drug cost measured at invoice price levels; rebates and discounts are not reflected.28 Second, a previous study based on U.S. data (Lichtenberg (2014c)) showed that about 25% of the cost of new drugs is offset by reduced expenditure on old drugs.29 Third, our estimates indicated that, if no drugs had been launched during 1986–2001, the average length of 2016 hospital stays would have been about 16% higher. This suggests that hospital expenditure might have been 16% higher. According to the Canadian Institute for Health Information (2017), hospital expenditure in 2016 was 51.30 billion USD (= 66.63 billion CAD at a 0.77 USD/CAD exchange rate), so hospital expenditure might have been 8.21 billion USD (= 16% * 51.30 billion USD) higher. The reduction in hospital expenditure due to shorter average length of stay may have been larger than the expenditure on the drugs responsible for shorter hospital stays.

6. Summary

In this study, we performed an econometric assessment of the role that pharmaceutical innovation—the introduction and use of new drugs—has played in reducing the burden of disease in Canada, by investigating whether diseases for which more new drugs were launched had larger subsequent reductions in disease burden. Since utilization of a drug reaches a peak about 12–14 years after it was launched, we allowed for considerable lags in the relationship between new drug launches and the burden of disease.

We analyzed the impact of new drug launches on a comprehensive measure of disease burden—the age-standardized disability-adjusted life-years lost (DALY) rate—and on its two components: the age-standardized years of life lost (YLL) and years lost to disability (YLD) rates. We also analyzed the impact of new drug launches on the number of hospital discharges and on the average length of hospital stays.

We found that the number of DALYs lost is significantly inversely related to the number of drugs that had ever been launched 9–20 years earlier, and that the number of YLLs is significantly inversely related to the number of drugs that had ever been launched 11–20 years earlier. The launch of a drug had the largest (most negative) impact on the number of DALYs and YLLs 15 years after it was launched.

The estimates indicated that if no drugs had been launched during 1986–2001, the age-standardized DALY rate would not have declined between 2000 and 2016; it might even have increased. Almost all (93%) of the reduction in DALYs was due to a reduction in YLL. The estimates implied that new drug launches during 1986–2001 reduced DALYs in 2016 by 21%, reduced YLLs in 2016 by 28%, and reduced YLDs in 2016 by 3%.

We estimated that drugs launched during 1986–2001 reduced the number of DALYs lost in 2016 by 2.31 million. Expenditure in 2016 on drugs launched during 1986–2001 per DALY gained in 2016 from those drugs was 2842 USD. Interventions that avert one DALY for less than average per capita income for a given country or region are generally considered to be very cost–effective; Canada's per capita GDP was 42,158 USD in 2016, so our estimates indicate that the new drugs launched during 1986–2001 were very cost–effective, overall.

Due to data limitations, we were unable to control for non-pharmaceutical medical innovations. Evidence from previous studies suggests that this is unlikely to cause significant bias in our estimates, because (1) the vast majority (88%) of private U.S. biomedical research funding came from pharmaceutical and biotechnology firms, and (2) non-pharmaceutical medical innovation does not appear to be positively correlated across diseases with pharmaceutical innovation. But, arguendo, suppose that drugs launched during 1986–2001 reduced the number of DALYs lost in 2016 by half as much as we estimated: by 1.155 million, instead of 2.31 million, and that the other half was due to new medical devices. Then 2016 expenditure on those drugs per 2016 DALY reduction would be twice as high as we estimated: $5684, instead of $2842. Even this higher figure would indicate that the new drugs launched during 1986–2001 were very cost–effective in 2016.

Moreover, 2842 USD may be an overestimate of the true net cost in 2016 per DALY of drugs launched during 1986–2001. A previous study based on U.S. data showed that about 25% of the cost of new drugs is offset by reduced expenditure on old drugs. Also, our estimates indicated that, if no drugs had been launched during 1986–2001, the average length of 2016 hospital stays would have been about 16% higher. The reduction in hospital expenditure due to shorter average length of stay may have been larger than the expenditure on the drugs responsible for shorter hospital stays.

Funding

This article is based on a research report commissioned and funded by Merck Canada Inc.

Appendix

Appendix Table 1.

Age-standardized rates (per 100,000 population) of Disability-Adjusted Life-Years lost (DALYs), Years of Life Lost (YLLs), and Years of Healthy Life Lost due to Disability (YLDs), by cause, Canada, 2000 and 2016.

WHO Global Health Estimates cause	DALY		YLL		YLD
	2000	2016	2000	2016	2000	2016
0 All Causes	25,087.3	21,463.0	14,818.3	11,121.2	10,269.0	10,341.9
10 Communicable, maternal, perinatal and nutritional conditions	1,181.7	1,142.7	824.4	757.2	357.3	385.5
20 Infectious and parasitic diseases	340.8	304.5	273.1	230.4	67.7	74.1
30 Tuberculosis	11.0	5.9	9.9	5.0	1.1	0.9
40 STDs excluding HIV	14.5	16.5	1.3	2.6	13.2	13.9
50 Syphilis	2.6	4.4	0.4	2.1	2.2	2.3
60 Chlamydia	2.1	2.1	0.2	0.1	1.9	2.0
70 Gonorrhoea	1.6	1.5	0.5	0.3	1.1	1.2
80 Trichomoniasis	3.3	3.4	0.0	0.0	3.3	3.4
85 Genital herpes	2.6	2.7	0.0	0.0	2.6	2.7
90 Other STDs	2.4	2.4	0.2	0.1	2.2	2.2
100 HIV/AIDS	107.9	59.5	85.0	32.4	22.9	27.1
110 Diarrhoeal diseases	26.6	66.2	13.4	50.9	13.3	15.3
120 Childhood-cluster diseases	2.1	1.2	1.2	0.4	0.9	0.8
130 Whooping cough	1.8	1.0	0.9	0.3	0.9	0.8
140 Diphtheria	0.0	0.0	0.0	0.0	0.0	0.0
150 Measles	0.2	0.1	0.2	0.1	0.0	0.0
160 Tetanus	0.0	0.0	0.0	0.0	0.0	0.0
170 Meningitis	19.8	11.8	16.7	8.4	3.1	3.4
180 Encephalitis	25.3	25.3	21.8	21.3	3.5	3.9
185 Hepatitis	22.0	11.4	20.3	10.7	1.7	0.6
186 Acute hepatitis A	5.2	2.0	4.5	1.7	0.8	0.3
190 Acute hepatitis B	7.5	5.7	6.7	5.4	0.8	0.3
200 Acute hepatitis C	0.0	0.0	0.0	0.0	0.0	0.0
205 Acute hepatitis E	9.2	3.7	9.1	3.7	0.1	0.0
210 Parasitic and vector diseases	2.2	0.9	1.6	0.3	0.7	0.5
220 Malaria	0.0	0.0	0.0	0.0	0.0	0.0
230 African trypanosomiasis	0.0	0.0	0.0	0.0	0.0	0.0
240 Chagas disease	0.2	0.2	0.0	0.0	0.2	0.2
250 Schistosomiasis	0.0	0.0	0.0	0.0	0.0	0.0
260 Leishmaniasis	0.0	0.0	0.0	0.0	0.0	0.0
270 Lymphatic filariasis	0.0	0.0	0.0	0.0	0.0	0.0
280 Onchocerciasis	0.0	0.0	0.0	0.0	0.0	0.0
285 Cysticercosis	1.1	0.1	0.9	0.0	0.2	0.1
295 Echinococcosis	0.9	0.3	0.6	0.0	0.2	0.3
300 Dengue	0.0	0.0	0.0	0.0	0.0	0.0
310 Trachoma	0.0	0.0	0.0	0.0	0.0	0.0
315 Yellow fever	0.0	0.0	0.0	0.0	0.0	0.0
320 Rabies	0.1	0.2	0.1	0.2	0.0	0.0
330 Intestinal nematode infections	0.0	0.0	0.0	0.0	0.0	0.0
340 Ascariasis	0.0	0.0	0.0	0.0	0.0	0.0
350 Trichuriasis	0.0	0.0	0.0	0.0	0.0	0.0
360 Hookworm disease	0.0	0.0	0.0	0.0	0.0	0.0
362 Food-bourne trematodes	0.0	0.0	0.0	0.0	0.0	0.0
365 Leprosy	0.0	0.0	0.0	0.0	0.0	0.0
370 Other infectious diseases	109.3	105.9	102.0	98.3	7.4	7.6
380 Respiratory infections	376.1	348.3	245.3	218.0	130.7	130.3
390 Lower respiratory infections	248.2	220.4	243.8	216.0	4.3	4.4
400 Upper respiratory infections	102.2	103.2	0.9	1.5	101.3	101.6
410 Otitis media	25.7	24.8	0.6	0.5	25.1	24.3
420 Maternal conditions	7.6	6.5	6.0	4.8	1.5	1.6
490 Neonatal conditions	383.5	415.2	278.4	287.7	105.1	127.5
500 Preterm birth complications	226.7	242.1	156.7	150.9	70.0	91.2
510 Birth asphyxia and birth trauma	75.7	68.7	59.3	50.4	16.4	18.3
520 Neonatal sepsis and infections	28.9	28.4	13.9	14.6	15.0	13.8
530 Other neonatal conditions	52.1	76.0	48.5	71.8	3.6	4.2
540 Nutritional deficiencies	73.8	68.3	21.5	16.3	52.2	52.0
550 Protein-energy malnutrition	12.9	11.8	8.5	6.6	4.4	5.1
560 Iodine deficiency	12.3	12.4	0.0	0.0	12.3	12.3
570 Vitamin A deficiency	0.0	0.0	0.0	0.0	0.0	0.0
580 Iron-deficiency anaemia	45.5	42.7	10.1	8.4	35.5	34.3
590 Other nutritional deficiencies	3.1	1.4	2.9	1.2	0.1	0.2
600 Noncommunicable diseases	21,373.3	18,102.5	12,347.7	9,052.0	9,025.7	9,050.6
610 Malignant neoplasms	4,885.6	3,750.2	4,734.1	3,601.0	151.5	149.2
620 Mouth and oropharynx cancers	87.3	74.7	84.1	72.0	3.2	2.7
621 Lip and oral cavity	44.9	39.2	42.6	37.4	2.3	1.8
622 Nasopharynx	10.9	8.0	10.6	7.8	0.3	0.2
623 Other pharynx	31.5	27.5	30.9	26.9	0.7	0.6
630 Oesophagus cancer	122.0	102.9	120.7	101.7	1.3	1.3
640 Stomach cancer	162.5	107.3	158.3	103.5	4.3	3.8
650 Colon and rectum cancers	536.3	427.1	516.5	408.5	19.8	18.6
660 Liver cancer	100.5	127.5	99.6	126.0	1.0	1.5
661 Liver cancer secondary to hepatitis B	16.6	21.0	16.5	20.8	0.1	0.1
662 Liver cancer secondary to hepatitis C	36.6	47.6	36.5	47.4	0.1	0.2
663 Liver cancer secondary to alcohol use	30.8	38.7	30.5	38.2	0.4	0.5
664 Other liver cancer	16.5	20.3	16.0	19.6	0.4	0.7
670 Pancreas cancer	222.9	204.8	220.7	202.6	2.3	2.2
680 Trachea, bronchus, lung cancers	1,261.9	943.2	1,247.8	930.0	14.1	13.2
690 Melanoma and other skin cancers	91.6	88.6	86.8	82.5	4.8	6.0
691 Malignant skin melanoma	78.8	72.2	74.3	66.8	4.5	5.3
692 Non-melanoma skin cancer	12.8	16.4	12.5	15.7	0.3	0.7
700 Breast cancer	484.8	326.9	456.5	300.9	28.4	26.0
710 Cervix uteri cancer	41.9	34.8	39.7	32.9	2.2	1.9
720 Corpus uteri cancer	60.7	60.4	56.8	55.8	3.8	4.6
730 Ovary cancer	115.5	91.4	112.3	88.6	3.2	2.8
740 Prostate cancer	244.3	159.9	220.6	138.8	23.7	21.1
742 Testicular cancer	7.5	7.8	6.6	6.7	0.9	1.1
745 Kidney cancer	117.3	93.3	113.8	89.8	3.5	3.4
750 Bladder cancer	107.7	89.3	101.5	83.3	6.2	6.0
751 Brain and nervous system cancers	183.6	166.4	181.2	163.6	2.4	2.8
752 Gallbladder and biliary tract cancer	38.9	24.2	37.9	23.5	1.0	0.7
753 Larynx cancer	44.7	22.2	43.2	21.0	1.5	1.2
754 Thyroid cancer	12.2	12.4	10.1	10.0	2.1	2.3
755 Mesothelioma	26.6	22.3	26.1	21.8	0.5	0.5
760 Lymphomas, multiple myeloma	336.5	229.0	326.0	216.8	10.5	12.2
761 Hodgkin lymphoma	18.2	12.5	17.0	11.1	1.2	1.4
762 Non-Hodgkin lymphoma	233.5	145.6	226.8	137.5	6.7	8.1
763 Multiple myeloma	84.9	70.9	82.2	68.1	2.7	2.8
770 Leukaemia	192.9	153.2	186.3	145.9	6.6	7.3
780 Other malignant neoplasms	285.4	180.6	281.1	174.8	4.3	5.8
790 Other neoplasms	107.5	75.8	98.7	64.3	8.8	11.5
800 Diabetes mellitus	851.9	699.9	414.5	282.8	437.5	417.1
810 Endocrine, blood, immune disorders	230.9	232.1	162.2	159.5	68.7	72.6
811 Thalassaemias	10.0	10.3	1.5	1.1	8.5	9.2
812 Sickle cell disorders and trait	4.0	4.2	1.6	1.7	2.4	2.5
813 Other haemoglobinopathies and haemolytic anaemias	19.2	13.4	17.1	11.6	2.1	1.8
814 Other endocrine, blood and immune disorders	197.7	204.2	142.0	145.1	55.8	59.1
820 Mental and substance use disorders	2,848.6	2,973.0	278.1	293.5	2,570.5	2,679.4
830 Depressive disorders	662.0	633.0	0.0	0.0	662.0	633.0
831 Major depressive disorder	499.3	469.0	0.0	0.0	499.3	469.0
832 Dysthymia	162.7	164.0	0.0	0.0	162.7	164.0
840 Bipolar disorder	176.4	171.3	1.3	1.1	175.1	170.3
850 Schizophrenia	226.9	225.3	8.8	4.3	218.1	220.9
860 Alcohol use disorders	280.5	272.5	94.1	77.4	186.4	195.1
870 Drug use disorders	536.6	627.1	164.8	204.7	371.7	422.4
871 Opioid use disorders	348.0	419.3	120.7	142.3	227.2	277.0
872 Cocaine use disorders	61.1	72.7	13.5	22.1	47.6	50.6
873 Amphetamine use disorders	19.4	23.6	4.7	7.8	14.6	15.8
874 Cannabis use disorders	31.0	24.2	0.0	0.0	31.0	24.2
875 Other drug use disorders	77.1	87.4	25.9	32.5	51.2	54.9
880 Anxiety disorders	482.4	513.5	0.0	0.0	482.4	513.5
890 Eating disorders	54.8	60.5	2.8	3.6	52.1	56.9
900 Autism and Asperger syndrome	138.9	140.4	0.0	0.0	138.9	140.4
910 Childhood behavioural disorders	79.9	78.9	0.0	0.0	79.9	78.9
911 Attention deficit/hyperactivity syndrome	12.3	12.2	0.0	0.0	12.3	12.2
912 Conduct disorder	67.5	66.7	0.0	0.0	67.5	66.7
920 Idiopathic intellectual disability	32.6	70.3	6.4	2.5	26.3	67.8
930 Other mental and behavioural disorders	177.7	180.3	0.0	0.0	177.7	180.3
940 Neurological conditions	1,596.6	1,761.2	706.3	874.6	890.3	886.5
950 Alzheimer disease and other dementias	501.6	708.7	371.5	559.2	130.1	149.5
960 Parkinson disease	90.9	95.8	65.8	70.0	25.2	25.8
970 Epilepsy	96.7	87.1	34.3	33.3	62.4	53.7
980 Multiple sclerosis	86.0	88.5	43.5	35.7	42.4	52.8
990 Migraine	507.9	480.4	0.0	0.0	507.9	480.4
1000 Non-migraine headache	102.2	99.5	0.0	0.0	102.2	99.5
1010 Other neurological conditions	211.2	201.2	191.1	176.4	20.1	24.8
1020 Sense organ diseases	893.3	911.0	0.6	0.6	892.7	910.5
1030 Glaucoma	8.0	8.0	0.1	0.0	8.0	8.0
1040 Cataracts	29.9	15.5	0.0	0.0	29.9	15.5
1050 Uncorrected refractive errors	161.3	164.2	0.0	0.0	161.3	164.2
1060 Macular degeneration	12.8	11.0	0.0	0.0	12.8	11.0
1070 Other vision loss	46.1	37.5	0.0	0.0	46.1	37.5
1080 Other hearing loss	530.4	569.3	0.0	0.0	530.4	569.2
1090 Other sense organ disorders	104.8	105.6	0.5	0.6	104.3	105.0
1100 Cardiovascular diseases	4,654.4	2,733.9	4,049.1	2,200.5	605.3	533.3
1110 Rheumatic heart disease	28.5	19.5	26.4	17.7	2.1	1.8
1120 Hypertensive heart disease	58.1	62.2	47.5	52.7	10.6	9.5
1130 Ischaemic heart disease	2,769.0	1,416.0	2,668.8	1,331.7	100.2	84.3
1140 Stroke	985.7	609.9	761.5	400.9	224.2	209.0
1141 Ischaemic stroke	639.4	391.5	450.1	220.2	189.3	171.3
1142 Haemorrhagic stroke	346.2	218.4	311.4	180.7	34.9	37.8
1150 Cardiomyopathy, myocarditis, endocarditis	130.7	103.3	114.8	89.5	15.9	13.8
1160 Other circulatory diseases	682.5	523.0	430.0	308.1	252.4	214.9
1170 Respiratory diseases	1,147.6	992.7	743.9	611.5	403.8	381.2
1180 Chronic obstructive pulmonary disease	664.4	540.8	583.6	455.8	80.8	85.0
1190 Asthma	337.8	296.9	27.0	15.1	310.7	281.9
1200 Other respiratory diseases	145.4	155.0	133.2	140.6	12.2	14.4
1210 Digestive diseases	637.6	592.5	553.6	504.6	84.0	88.0
1220 Peptic ulcer disease	37.5	33.5	24.5	19.0	13.0	14.5
1230 Cirrhosis of the liver	232.1	251.8	215.3	234.5	16.8	17.3
1231 Cirrhosis due to hepatitis B	30.4	34.0	29.2	32.8	1.2	1.2
1232 Cirrhosis due to hepatitis C	54.5	57.1	50.0	52.5	4.5	4.6
1233 Cirrhosis due to alcohol use	98.7	106.1	92.5	99.6	6.3	6.4
1234 Other liver cirrhosis	48.6	54.7	43.6	49.6	4.9	5.1
1240 Appendicitis	3.8	3.3	2.7	2.2	1.1	1.1
1241 Gastritis and duodenitis	19.5	20.2	3.8	3.0	15.6	17.2
1242 Paralytic ileus and intestinal obstruction	32.0	31.3	30.9	30.1	1.1	1.2
1244 Inflammatory bowel disease	27.6	18.1	18.2	8.8	9.4	9.4
1246 Gallbladder and biliary diseases	27.2	25.9	22.7	21.1	4.6	4.8
1248 Pancreatitis	29.6	25.2	25.9	21.3	3.7	3.9
1250 Other digestive diseases	228.3	183.1	209.6	164.6	18.7	18.6
1260 Genitourinary diseases	502.7	447.1	217.4	159.8	285.2	287.3
1270 Kidney diseases	284.6	218.3	179.8	113.8	104.8	104.5
1271 Acute glomerulonephritis	0.4	0.3	0.4	0.3	0.0	0.0
1272 Chronic kidney disease due to diabetes	141.1	103.6	87.1	50.7	54.0	52.9
1273 Other chronic kidney disease	143.0	114.4	92.3	62.8	50.7	51.6
1280 Benign prostatic hyperplasia	61.7	63.5	1.9	2.7	59.8	60.8
1290 Urolithiasis	5.9	6.5	1.4	1.9	4.5	4.6
1300 Other urinary diseases	40.9	47.3	33.9	40.7	6.9	6.6
1310 Infertility	9.7	9.6	0.0	0.0	9.7	9.6
1320 Gynecological diseases	100.0	101.9	0.4	0.7	99.5	101.2
1330 Skin diseases	308.4	313.7	15.6	17.1	292.8	296.6
1340 Musculoskeletal diseases	1,964.1	1,950.8	83.4	70.7	1,880.7	1,880.1
1350 Rheumatoid arthritis	105.3	116.8	16.5	11.2	88.9	105.6
1360 Osteoarthritis	245.4	251.8	6.1	4.2	239.4	247.7
1370 Gout	30.2	30.8	0.4	0.5	29.9	30.4
1380 Back and neck pain	893.5	872.5	2.2	3.3	891.3	869.2
1390 Other musculoskeletal disorders	689.7	678.9	58.3	51.7	631.3	627.2
1400 Congenital anomalies	366.4	307.3	254.6	192.5	111.8	114.7
1410 Neural tube defects	25.9	22.3	12.2	8.0	13.7	14.3
1420 Cleft lip and cleft palate	0.9	0.8	0.1	0.1	0.8	0.8
1430 Down syndrome	23.8	28.5	16.1	20.8	7.7	7.7
1440 Congenital heart anomalies	115.3	73.3	102.0	59.9	13.3	13.5
1450 Other chromosomal anomalies	45.3	47.5	32.8	35.3	12.5	12.2
1460 Other congenital anomalies	155.2	134.8	91.4	68.6	63.8	66.2
1470 Oral conditions	342.7	343.6	0.6	1.3	342.1	342.4
1480 Dental caries	27.1	26.9	0.0	0.0	27.1	26.9
1490 Periodontal disease	82.8	83.6	0.1	0.1	82.7	83.6
1500 Edentulism	175.7	175.0	0.1	0.1	175.6	174.9
1502 Other oral disorders	57.1	58.1	0.5	1.1	56.6	56.9
1505 Sudden infant death syndrome	35.0	17.6	35.0	17.6	0.0	0.0

Appendix Table 2.

Number of (WHO ATC5) chemical substances ever launched, by cause, Canada, 1980–2016.

WHO Global Health Estimates cause	1980	1986	1992	1998	2004	2010	2016
All chemical substances	828	1033	1276	1614	1825	2000	2232
30 Tuberculosis	10	10	10	15	15	15	15
50 Syphilis	4	4	4	5	5	5	5
60 Chlamydia	5	5	5	5	5	5	5
70 Gonorrhoea	6	7	10	10	10	11	11
80 Trichomoniasis	2	3	3	3	3	3	3
85 Genital herpes	0	2	2	4	4	4	4
90 Other STDs	1	2	2	2	3	4	4
100 HIV/AIDS	0	1	5	15	25	34	40
110 Diarrhoeal diseases	15	17	20	22	23	25	26
130 Whooping cough	3	4	4	5	5	5	5
140 Diphtheria	4	6	6	7	7	7	7
150 Measles	2	4	4	5	5	5	5
160 Tetanus	8	10	10	11	11	11	11
170 Meningitis	10	17	20	22	25	27	28
180 Encephalitis	0	0	0	1	1	2	2
186 Acute hepatitis A	1	1	1	3	4	4	4
190 Acute hepatitis B	1	3	5	7	10	12	12
200 Acute hepatitis C	0	2	3	3	6	6	15
220 Malaria	5	5	6	9	9	9	9
230 African trypanosomiasis	0	0	1	1	1	1	1
250 Schistosomiasis	0	0	0	1	1	1	1
260 Leishmaniasis	1	1	2	2	2	2	2
270 Lymphatic filariasis	0	1	1	1	1	1	1
280 Onchocerciasis	0	1	1	1	1	1	1
295 Echinococcosis	0	1	1	1	1	1	1
310 Trachoma	1	1	1	1	1	1	1
315 Yellow fever	1	1	1	1	1	1	1
320 Rabies	0	2	2	2	2	2	2
340 Ascariasis	2	2	3	3	3	3	3
360 Hookworm disease	1	1	2	2	2	2	2
362 Food-bourne trematodes	3	3	3	4	4	4	4
365 Leprosy	2	2	2	2	2	3	3
370 Other infectious diseases	88	109	127	158	169	177	177
390 Lower respiratory infections	37	45	56	62	70	75	75
400 Upper respiratory infections	47	53	62	68	70	72	73
410 Otitis media	18	20	25	25	26	27	27
420 Maternal conditions	21	27	34	37	38	39	40
500 Preterm birth complications	4	4	8	8	8	9	9
530 Other neonatal conditions	11	12	14	14	14	14	14
550 Protein-energy malnutrition	2	2	2	2	3	3	3
560 Iodine deficiency	4	4	4	5	5	5	5
570 Vitamin A deficiency	8	9	9	9	9	9	9
580 Iron-deficiency anaemia	5	5	6	6	7	10	10
590 Other nutritional deficiencies	39	42	45	47	48	49	49
620 Mouth and oropharynx cancers	1	3	3	4	4	4	4
630 Oesophagus cancer	1	4	4	6	6	6	6
640 Stomach cancer	2	4	5	7	8	8	9
650 Colon and rectum cancers	3	3	3	6	6	12	15
660 Liver cancer	0	1	1	2	2	3	3
670 Pancreas cancer	1	3	5	6	6	9	10
680 Trachea, bronchus, lung cancers	5	11	13	19	21	23	30
691 Malignant skin melanoma	3	4	5	5	5	5	12
692 Non-melanoma skin cancer	0	0	0	0	1	1	1
700 Breast cancer	8	14	18	28	32	35	40
710 Cervix uteri cancer	1	3	4	5	5	8	8
720 Corpus uteri cancer	0	1	1	1	1	1	1
730 Ovary cancer	6	10	13	17	17	19	20
740 Prostate cancer	2	5	8	12	13	15	19
742 Testicular cancer	3	5	7	7	7	7	7
745 Kidney cancer	1	2	4	4	4	6	7
750 Bladder cancer	6	9	11	12	12	12	12
751 Brain and nervous system cancers	11	12	12	12	13	13	14
752 Gallbladder and biliary tract cancer	0	0	0	1	1	1	1
754 Thyroid cancer	0	0	0	0	0	1	3
755 Mesothelioma	2	3	4	4	4	6	6
761 Hodgkin lymphoma	11	15	17	17	17	17	20
762 Non-Hodgkin lymphoma	16	24	28	28	29	34	42
763 Multiple myeloma	9	13	15	15	15	18	24
770 Leukaemia	16	23	28	32	35	41	52
800 Diabetes mellitus	7	8	13	24	35	45	48
811 Thalassaemias	0	0	0	0	0	1	2
812 Sickle cell disorders and trait	2	2	2	2	2	2	2
813 Other haemoglobinopathies and haemolytic anaemias	8	8	9	9	11	15	15
814 Other endocrine, blood and immune disorders	84	101	125	149	179	199	216
830 Depressive disorders	10	13	19	23	26	27	28
840 Bipolar disorder	6	7	7	11	11	12	13
850 Schizophrenia	15	16	17	22	22	25	25
860 Alcohol use disorders	11	13	14	14	14	15	15
871 Opioid use disorders	0	0	1	2	3	4	4
880 Anxiety disorders	16	19	24	26	28	30	30
890 Eating disorders	7	8	11	11	11	11	11
900 Autism and Asperger syndrome	2	2	2	2	2	2	2
911 Attention deficit/hyperactivity syndrome	1	2	2	2	3	3	3
912 Conduct disorder	0	0	0	1	1	1	1
920 Idiopathic intellectual disability	0	0	0	1	1	1	1
930 Other mental and behavioural disorders	27	30	37	49	57	60	61
950 Alzheimer disease and other dementias	0	0	0	0	1	1	1
960 Parkinson disease	7	8	11	15	17	20	21
970 Epilepsy	15	15	16	21	23	26	29
980 Multiple sclerosis	7	9	12	16	17	19	24
990 Migraine	13	13	15	18	22	22	22
1000 Non-migraine headache	8	9	10	10	11	12	12
1010 Other neurological conditions	28	31	35	45	49	53	55
1030 Glaucoma	7	10	13	18	20	20	22
1040 Cataracts	1	1	1	1	1	1	1
1050 Uncorrected refractive errors	2	2	3	3	4	5	5
1060 Macular degeneration	0	0	0	0	1	4	5
1070 Other vision loss	11	13	13	14	16	19	20
1080 Other hearing loss	1	1	1	1	1	1	1
1090 Other sense organ disorders	45	52	61	73	79	82	83
1110 Rheumatic heart disease	12	12	12	14	14	14	14
1120 Hypertensive heart disease	22	31	44	60	69	75	80
1130 Ischaemic heart disease	17	24	32	45	52	55	58
1140 Stroke	4	6	10	14	15	17	19
1150 Cardiomyopathy, myocarditis, endocarditis	18	23	26	26	26	27	27
1160 Other circulatory diseases	59	75	92	115	128	137	145
1180 Chronic obstructive pulmonary disease	33	36	47	55	60	62	69
1190 Asthma	19	21	27	33	38	39	40
1200 Other respiratory diseases	44	53	72	80	84	87	94
1220 Peptic ulcer disease	4	8	13	16	19	19	19
1230 Cirrhosis of the liver	5	6	7	8	12	14	19
1241 Gastritis and duodenitis	0	0	1	1	1	1	1
1242 Paralytic ileus and intestinal obstruction	1	2	2	2	2	2	2
1244 Inflammatory bowel disease	8	9	11	13	15	16	17
1246 Gallbladder and biliary diseases	5	5	6	8	8	8	8
1248 Pancreatitis	4	4	5	5	5	5	5
1250 Other digestive diseases	58	71	88	102	104	106	112
1270 Kidney diseases	26	34	44	54	59	63	66
1280 Benign prostatic hyperplasia	1	2	5	6	9	9	11
1290 Urolithiasis	7	7	7	7	7	7	7
1300 Other urinary diseases	40	49	60	73	78	82	86
1310 Infertility	5	5	7	10	11	11	12
1320 Gynecological diseases	22	27	29	31	36	39	43
1330 Skin diseases	94	111	132	149	171	181	191
1350 Rheumatoid arthritis	20	24	25	28	38	43	44
1360 Osteoarthritis	15	18	19	21	25	25	25
1370 Gout	6	6	6	6	6	7	7
1380 Back and neck pain	19	22	24	28	34	36	37
1390 Other musculoskeletal disorders	57	68	79	91	105	115	118
1410 Neural tube defects	1	1	1	1	1	1	1
1440 Congenital heart anomalies	0	0	0	0	1	1	2
1450 Other chromosomal anomalies	0	0	1	1	1	1	1
1460 Other congenital anomalies	7	9	11	12	12	13	15
1480 Dental caries	4	4	5	5	5	5	5
1490 Periodontal disease	9	10	13	14	14	15	15
1502 Other oral disorders	16	20	23	26	26	27	27

Appendix Table 3.

Number of hospital discharges and average length of stay (in days), by cause, Canada, 2000 and 2016.

Cause	Number of hospital discharges		Average length of stay
	2000	2016	2000	2016
0000 All causes	28,85,062	30,57,503	7.2	8.1
0101 Intestinal infectious diseases except diarrhoea	14,131	16,801	4.8	7.1
0103 Tuberculosis	1,015	1,039	20.2	21.8
0104 Septicaemia	9,230	26,590	11.1	12.2
0105 Human immunodeficiency virus [HIV] disease	1,652	944	13.4	16.4
0106 Other infectious and parasitic diseases	16,522	20,835	6.2	8.1
0200 Neoplasms	2,15,414	2,02,310	9.6	8.4
0201 Malignant neoplasm of colon, rectum and anus	22,081	22,347	13.7	10.3
0202 Malignant neoplasms of trachea, bronchus and lung	22,204	18,851	12.2	10.7
0203 Malignant neoplasms of skin	1,991	2,012	6.3	6.9
0204 Malignant neoplasm of breast	17,796	9,822	4.5	4.2
0205 Malignant neoplasm of uterus	5,730	7,424	6.4	4.3
0206 Malignant neoplasm of ovary	3,270	3,021	11.0	7.8
0207 Malignant neoplasm of prostate	12,485	10,491	7.1	4.7
0208 Malignant neoplasm of bladder	10,214	8,281	6.0	6.2
0209 Other malignant neoplasms	74,078	82,082	13.0	11.0
0210 Carcinoma in situ	3,561	3,555	3.7	3.2
0211 Benign neoplasm of colon, rectum and anus	2,253	3,046	7.4	5.7
0212 Leiomyoma of uterus	18,514	9,934	3.6	2.1
0213 Other benign neoplasms and neoplasms of uncertain or unknown behaviour	21,237	21,444	5.9	5.2
0300 Diseases of the blood and bloodforming organs and certain disorders involving the immune mechanism	25,688	27,984	6.4	6.6
0301 Anaemias	13,960	14,909	6.8	6.5
0302 Other diseases of the blood and bloodforming organs and certain disorders involving the immune mechanism	11,728	13,075	6.0	6.7
0400 Endocrine, nutritional and metabolic diseases	64,573	78,580	8.2	7.8
0401 Diabetes mellitus	30,021	37,127	10.0	9.9
0402 Other endocrine, nutritional and metabolic diseases	34,552	41,453	6.7	5.9
0600 Diseases of the nervous system	43,652	59,431	11.6	14.0
0602 Multiple sclerosis	1,856	1,425	15.7	16.5
0603 Epilepsy	7,889	13,077	5.7	6.5
0604 Transient cerebral ischaemic attacks and related syndromes	10,970	9,057	5.7	4.5
0605 Other diseases of the nervous system	20,145	30,396	13.6	14.4
0700 Diseases of the eye and adnexa	15,436	5,898	2.3	3.3
0701 Cataract	1,662	220	2.0	2.0
0702 Other diseases of the eye and adnexa	13,774	5,678	2.3	3.3
0800 Diseases of the ear and mastoid process	11,902	8,534	2.7	3.1
0900 Diseases of the circulatory system	4,34,532	3,80,737	8.7	8.0
0901 Hypertensive diseases	12,334	6,594	8.5	5.6
0903 Acute myocardial infarction	64,920	71,909	8.4	5.2
0905 Pulmonary heart disease & diseases of pulmonary circulation	7,725	12,254	9.2	7.5
0906 Conduction disorders and cardiac arrhythmias	54,463	53,872	5.1	5.2
0907 Heart failure	60,321	65,510	10.1	10.1
0908 Cerebrovascular diseases	52,954	53,049	16.7	13.2
0909 Atherosclerosis	7,135	6,831	11.9	10.8
0910 Varicose veins of lower extremities	2,404	1,046	5.7	11.7
0911 Other diseases of the circulatory system	49,294	55,732	9.1	8.7
1000 Diseases of the respiratory system	2,57,572	2,72,315	6.6	7.4
1001 Acute upper respiratory infections and influenza	17,875	22,704	2.7	4.6
1002 Pneumonia	78,938	67,974	7.9	7.4
1003 Other acute lower respiratory infections	20,215	17,119	3.7	3.7
1004 Chronic diseases of tonsils and adenoids	12,283	5,858	1.2	1.1
1005 Other diseases of upper respiratory tract	11,038	8,152	2.4	2.7
1006 Chronic obstructive pulmonary disease and bronchiectasis	55,757	89,897	9.2	8.1
1007 Asthma	31,010	11,443	3.5	2.9
1008 Other diseases of the respiratory system	30,456	49,168	9.9	11.4
1101 Disorders of teeth and supporting structures	6,768	5,826	2.2	1.8
1102 Other diseases of oral cavity, salivary glands and jaws	2,733	2,893	3.7	4.5
1103 Diseases of oesophagus	10,681	7,877	5.2	5.6
1104 Peptic ulcer	11,149	9,692	7.3	7.3
1105 Dyspepsia and other diseases of stomach and duodenum	13,657	7,378	4.9	6.4
1106 Diseases of appendix	29,737	38,581	3.6	2.4
1107 Inguinal hernia	20,702	10,714	2.5	3.2
1108 Other abdominal hernia	17,446	17,985	4.3	4.5
1109 Crohn's disease and ulcerative colitis	12,586	10,099	9.2	7.8
1111 Paralytic ileus and intestinal obstruction without hernia	25,374	30,123	7.5	7.0
1112 Diverticular disease of intestine	19,127	18,358	7.3	6.2
1113 Diseases of anus and rectum	8,512	9,263	4.2	4.4
1114 Other diseases of intestine	12,731	14,515	8.3	8.4
1115 Alcoholic liver disease	4,755	6,185	11.9	12.5
1116 Other diseases of liver	6,802	9,911	11.5	11.2
1117 Cholelithiasis	49,080	33,606	3.7	4.0
1118 Other diseases of gall bladder and biliary tract	10,080	13,067	5.7	6.0
1119 Diseases of pancreas	16,037	24,021	8.3	6.2
1120 Other diseases of the digestive system	18,337	25,781	7.3	6.9
1200 Diseases of the skin and subcutaneous tissue	31,043	33,613	7.9	9.3
1201 Infections of the skin and subcutaneous tissue	23,478	26,634	6.6	7.9
1202 Dermatitis, eczema and papulosquamous disorders	1,616	1,744	5.4	5.7
1203 Other diseases of the skin and subcutaneous tissue	5,949	5,235	13.7	17.2
1300 Diseases of the musculoskeletal system and connective tissue	1,30,720	1,87,483	7.0	5.7
1301 Coxarthrosis [arthrosis of hip]	.	36,682	.	3.6
1302 Gonarthrosis [arthrosis of knee]	.	62,850	.	3.5
1303 Internal derangement of knee	4,123	1,028	1.7	1.6
1305 Systemic connective tissue disorders	3,699	3,337	12.2	10.7
1306 Deforming dorsopathies and spondylopathies	7,873	15,625	8.8	8.4
1307 Intervertebral disc disorders	15,726	10,846	5.3	5.4
1308 Dorsalgia	7,297	7,776	6.3	7.6
1309 Soft tissue disorders	13,985	12,252	4.2	9.4
1310 Other disorders of the musculoskeletal system and connective tissue	23,489	16,805	9.0	11.7
1400 Diseases of the genitourinary system	1,77,274	1,64,458	4.3	5.2
1401 Glomerular and renal tubulo-interstitial diseases	16,731	24,209	5.2	4.9
1402 Renal failure	11,723	23,905	12.9	10.0
1403 Urolithiasis	25,438	15,087	2.7	2.4
1404 Other diseases of the urinary system	30,805	40,709	5.4	7.6
1405 Hyperplasia of prostate	15,512	13,495	3.5	2.5
1406 Other diseases of male genital organs	5,424	4,207	3.2	4.5
1407 Disorders of breast	10,073	2,459	1.5	1.7
1408 Inflammatory diseases of female pelvic organs	6,345	3,589	3.5	3.8
1409 Menstrual, menopausal and other female genital conditions	19,072	13,015	3.0	1.8

Appendix Table 4.

Estimates of β_k parameters of semi-logarithmic version of eq. (4), Δln(Y_d) = β_k (CUM_DRUG_d,2016-k - CUM_DRUG_d,2000-k) + δ’+ ε_d’

row	lag	Estimate	Std. Err.	t Value	p-value	N	mean (regressor)	βk * mean (regressor)
1	0	−0.0056	0.003	−2.04	0.044	136	8.357	−0.047
2	1	−0.0026	0.003	−0.99	0.324	136	8.682	−0.022
3	2	−0.0037	0.002	−1.48	0.142	136	9.203	−0.034
4	3	−0.0048	0.002	−2.03	0.044	136	9.811	−0.047
5	4	−0.0055	0.002	−2.62	0.010	136	11.028	−0.061
6	5	−0.0054	0.002	−2.66	0.009	136	11.464	−0.062
7	6	−0.0047	0.002	−2.32	0.022	136	11.816	−0.055
8	7	−0.0051	0.002	−2.66	0.009	136	12.535	−0.064
9	8	−0.0052	0.002	−2.79	0.006	136	12.814	−0.066
10	9	−0.0045	0.002	−2.65	0.009	136	14.021	−0.064
11	10	−0.0042	0.002	−2.46	0.016	136	14.097	−0.059
12	11	−0.0047	0.002	−2.83	0.005	136	14.691	−0.069
13	12	−0.0043	0.002	−2.60	0.011	136	14.678	−0.063
14	13	−0.0052	0.002	−3.23	0.002	136	14.769	−0.076
15	14	−0.0052	0.002	−3.35	0.001	136	14.923	−0.078
16	15	−0.0049	0.002	−3.14	0.002	136	14.972	−0.074
17	16	−0.0054	0.002	−3.29	0.001	136	14.877	−0.080
18	17	−0.0059	0.002	−3.65	0.000	136	14.770	−0.088
19	18	−0.0057	0.002	−3.39	0.001	136	14.121	−0.080
20	19	−0.0055	0.002	−3.26	0.001	136	13.946	−0.076
21	20	−0.0051	0.002	−2.87	0.005	136	13.276	−0.068

Appendix Fig. 1.

Age-standardized percentage reporting a diagnosis of diabetes, by sex, population aged 12 or older, canada, 2000/2001,2003,2005,2007, and 2008.1