Outpatient oral anticancer agent utilization and costs in Manitoba from 2003 to 2016: a population-based study

Abstract

Intervention

In April 2012, the Manitoba Home Cancer Drug Program (HCDP) was introduced to allow 100% coverage for eligible oral anticancer agents (OAA) and supportive medications for Manitobans with cancer requiring these therapies.

Research questions

What is the extent of use and cost of OAAs among outpatients in Manitoba from 2003/04 to 2015/16? Did the HCDP change OAA user and prescription patterns?

Methods

This was a retrospective, population-based study using administrative data to measure the prevalence of drug utilization over time and the impact of HCDP on OAA use and prescriptions using generalized linear models. Manitobans with cancer who filled an OAA or supportive medication covered by HCDP from 2003/04 to 2015/16 were included.

Results

This study included 22,393 people with cancer who filled an OAA prescription. The prevalence of OAA use increased from 222 per 100,000 to 328 per 100,000 from 2003/04 to 2015/16. Hormone therapy for breast cancer was the most common class of OAA used (increased from 154 per 100,000 to 231 per 100,000). We observed a 2.6-fold decrease in the prevalence of oral alkylating agents and a 10.7-fold increase in the prevalence of protein kinase inhibitors during the study period. The total cost of targeted OAAs per year for all Manitobans with cancer increased from $1.8 million to $19 million.

Conclusion

We observed an increase in OAA prevalence and the cost of oral targeted chemotherapy is high. Our findings underline the need for addressing these high-cost medications in future developments of a national drug program.

Supplementary Information

The online version contains supplementary material available at 10.17269/s41997-020-00464-6.

Introduction

Approximately 45% of Canadians are expected to develop cancer during their lifetime, and many of these individuals are anticipated to receive anticancer treatment (Canadian Cancer Society, 2019). For most types of cancers, anticancer medications have traditionally been comprised of a combination of injectable chemotherapy (Canadian Association of Pharmacy in Oncology (CAPhO), 2009). However, there has been a rapid growth in the number of oral anticancer agents (OAAs) available for use in Canada (Aisner, 2007). These OAAs, which include hormone therapy, targeted therapy, and traditional chemotherapy, have the potential to improve the convenience and quality of life for cancer patients during therapy (Canadian Association of Pharmacy in Oncology (CAPhO), 2009, Aisner, 2007). As compared with injectable chemotherapy, oral formulations do not require invasive administration and do not require the patient to come into clinic for administration by an oncology nurse (Aisner, 2007). However, patients are also held more accountable for ensuring the proper use and handling of these agents (Abbott et al., 2014, Abbott et al., 2011, Weingart et al., 2007, Leveque et al., 2014, Conde-Estevez et al., 2012, Timmers et al., 2014, Bassan et al., 2014, Hammond et al., 2012). Moreover, many of these oral formulations are costly (Timmers et al., 2012, Bennette et al., 2016, Dusetzina et al., 2019).

Provinces across Canada are aiming to provide drug coverage for both injectable and oral forms of anticancer therapies. However, the rising cost of OAAs, and in particular, targeted therapies, makes it difficult for public drug plans to balance the accessibility of essential medications for the population. In Manitoba, eligible prescription medications have been historically covered by the Manitoba Pharmacare program after an income-based deductible is paid by the patient each fiscal year. In April 2012, the Manitoba Home Cancer Drug Program (HCDP) was introduced in which eligible OAAs and supportive medications were 100% covered for Manitoba residents with cancer requiring these therapies. Patients who do not have drug coverage elsewhere, such as private or federal programs, are eligible for the HCDP. We aimed to examine the change in OAA utilization and cost over time in Manitoba, Canada, using comprehensive population-based data. We also aimed to describe the patient demographic using OAAs and the annual rate of contacts with the healthcare system among individuals with cancer using these agents for chemotherapy. Given the increased availability of these agents in the past twenty years, we hypothesized that there will be an increase in OAA use and expenditure on these agents over time. We anticipate that these findings will help inform future public drug policy planning, including a national drug strategy for managing high-cost medications.

Methods

Data sources

Data for this study were obtained from the Manitoba Population Research Data Repository located at the Manitoba Centre for Health Policy (MCHP). This Repository includes over 70 databases with information on health, education, social, justice, and registries linked using an encrypted identification number for almost all 1.3 million Manitoba residents. The datasets used in this study have been validated with excellent linkage accuracy and have been used extensively for research (Roos and Nicol, 1999, Roos et al., 2005, Mustard et al., 1999). The Drug Program Information Network (DPIN) database was used for information on prescription drug usage (date of prescription, drug name, quantity dispensed, prescriber, payer) of individuals living in Manitoba with the exception of medications received in the hospital setting and by First Nations patients receiving care from nursing stations. The Manitoba HCDP database was used to obtain information on dispensed prescriptions covered by this program. Medical services (physician billings) and hospitalization (hospital abstracts) files were used to capture contacts with the healthcare system and diagnoses identified by the International Classification of Diseases, Clinical Modification (ICD-9-CM or ICD-10-CA equivalent) codes. The Manitoba Health Insurance Registry and Statistics Canada census files provided the demographic information (e.g., age, sex) at the beginning of each interval. Vital statistics provided information on the date and cause of death, if applicable. The Consumer Price Index data from Statistics Canada were also used to determine cost of medications. This study was approved by the Human Research Ethics Board of the University of Manitoba and the Manitoba Health Seniors and Active Living (MHSAL) Health Information Privacy Committee.

Study population

All individuals living in Manitoba with at least one day of MHSAL coverage in the fiscal year was used as the population denominator. Manitobans who received at least one prescription for an OAA or specific supportive therapies within each quarter from April 1, 2003 through March 31, 2016 were identified as the population of OAA or supportive care medication users. We further identified all OAA users with a cancer diagnosis defined by a hospitalization or physician visit code for a specific cancer diagnosis (ICD-9-CM 140-209 except 173 or ICD-10-CA equivalent C00-C96 except for C44) (National Cancer Institute, 2014, Shih et al., 2015) in the prior or current fiscal year when the person filled a prescription for an OAA. Once an OAA user was determined to have a cancer diagnosis, we considered them to have had a cancer diagnosis for the study duration.

Medication exposure

All oral formulations of an anticancer drug that were available in Canada during the study period were identified and classified according to the Anatomical Therapeutic Chemical (ATC) classification, which divides the medications according to the organ system on which they act and the therapeutic or chemical characteristics of the medication. These agents were also grouped into broader categories as follows: (1) traditional agents which act on all rapidly dividing cells; (2) targeted agents, which interfere with specific molecules that are involved in the growth, progression, and spread of cancer; (3) hormonal agents, which are used for hormone responsive breast or prostate cancer (Table 1) (National Cancer Institute, 2014, Shih et al., 2015). We reported on supportive therapies separately. These mediations include drugs used to prevent or manage side effects as a result of chemotherapy (e.g., nausea/vomiting) and other medications that are typically used for non-cancer indications but are used off-label for their anticancer properties and are covered by the HCDP (Table 2).

Table 1.

Oral anticancer agent drug list

Drug group (National Cancer Institute)	Drug category	Specific drugs	ATC code
Traditional	Alkylating agents	Busulfan	L01A
		Chlorambucil	L01A
		Cyclophosphamide	L01A
		Lomustine	L01A
		Melphalan	L01A
		Mitotane	L01A
		Procarbazine	L01A
		Temozolomide	L01A
	Antimetabolites	Capecitabine	L01B
		Fludarabine phosphate	L01B
		Mercaptopurine	L01B
		Thioguanine	L01B
	Plant alkaloids	Etoposide	L01CB
	Other older anticancer agents	Hydroxyurea	L01XX
		Tretinoin	L01XX
	Lenalidomide and others	Lenalidomide	L04AX
		Pomalidomide	L04AX
		Thalidomide	L04AX
Targeted*	Protein kinase inhibitors for haematological malignancies	Bosutinib	L01XE
		Dasatinib	L01XE
		Imatinib mesylate	L01XE
		Nilotinib	L01XE
		Ruxolitinib	L01XE
	Protein kinase inhibitors for solid tumours	Afatanib	L01XE
		Axitinib	L01XE
		Crizotinib	L01XE
		Dabrafenib	L01XE
		Erlotinib	L01XE
		Gefitnib	L01XE
		Ibrutinib	L01XE
		Lapatinib	L01XE
		Pazopanib	L01XE
		Regorafenib	L01XE
		Sorafenib	L01XE
		Sunitinib	L01XE
		Trametinib	L01XE
		Vandetanib	L01XE
		Vemurafenib	L01XE
	Other newer anticancer agents	Everolimus	L01XE, L04AA
		Vismodegib	L01XX
Hormonal	Hormone therapy for breast cancer	Anastrozole	L02BG
		Exemestane	L02BG
		Letrozole	L02BG
		Tamoxifen	L02BA
	Hormone therapy for prostate cancer	Abiraterone	L02BX
		Bicalutamide	L02BB
		Enzalutamide	L02BB
		Flutamide	L02BB
		Nilutamide	L02BB

These drugs were also covered by HCDP at the start of the program

There are no prescriptions for thalidomide or fludarabine as they are not included in DPIN

*Targeted anticancer agent definition reference: http//www.cancer.gov/cancertopics/factsheet/Therapy/targeted

Reference: Raymond C, Leong C, Fransoo R, et al. Outpatient Oral Anticancer Agents in Manitoba. Winnipeg, MB. Manitoba Centre for Health Policy, Summer 2018

Table 2.

Supportive and other medications covered by the Home Cancer Drug Program

Drug category	Specific drugs	ATC code
Supportive	Estradiol	G03CA
	Megestrol	G03AC, G03DB, L02AB
	Oxycodone	N02AA
	Raloxifene	N02AA
Disease treatment	Acitretin	D05BB
	Anagrelide	L01XX
	Celecoxib	L01XX, M01AH
	Deferasirox	V03AC
	Imiquimod	D06BB
	Isotretinoin	D10AD, D10BA
	Ketoconazole	D01AC, G01AF, J02AB
	Methotrexate	L01BA, L04AX
Antinauseants	Aprepitant	A04AD
	Domperidone	A03FA
	Granisetron	A04AA
	Metoclopramide	A03FA
	Nabilone	A04AD
	Olanzapine	N05AH
	Ondansetron	A04AA
	Prochlorperazine	N05AB
Steroids	Dexamethasone	H02AB
	Hydrocortisone	H02AB
	Prednisone	A07EA, H02AB

Reference: Raymond C, Leong C, Fransoo R, et al. Outpatient Oral Anticancer Agents in Manitoba. Winnipeg, MB. Manitoba Centre for Health Policy, Summer 2018

Statistical analyses

The primary analysis examined the prevalent users for each medication or medication group, which was reported in quarterly intervals from April 1, 2003 through March 31, 2016. Prevalent use was expressed as users per 100,000 residents with cancer per quarter. The impact of the implementation of the HCDP Program on OAA medication use and prescriptions over time was evaluated with the generalized linear models for OAA users and OAA prescriptions for users with cancer. The time period before and after the introduction of the HCDP program was compared by evaluating the quarterly variation in rates of users and prescriptions in the lowest quarter of the year with the highest quarter of the year for OAA users with cancer. Using PROC GENMOD with a Poisson distribution, we included a contrast statement to determine whether the difference between the highest and lowest quarterly rates before the HCDP was implemented was significantly different (p ≤ 0.05) than after the program. Fiscal year was considered to be a categorical variable in the model.

The secondary analyses evaluated the annual prevalence of hospitalizations and physician visits among OAA users with a cancer diagnosis in the first year of receiving OAA. The cost per prescription (total cost, ingredient cost, and professional fee) was also determined per year. The total prescription cost per user per days supplied (sum of total cost divided by sum of total days supplied in a year) was determined per year. All costs were adjusted for inflation using the Consumer Price Index (CPI); all dollar values are in 2015 dollars (Statistics Canada, 2016). This analysis was performed for the OAA and supportive care medication users with a cancer diagnosis in the year of or year prior to their first prescription of interest. Total prescription cost was determined by the sum of ingredient cost and the professional dispensing fee. The ingredient cost is the cost of the medication without the pharmacy markup (Metge et al., 1999). The professional fee is the total fees charged on a per prescription basis by pharmacists for prescriptions dispensed.

Summary statistics were used to describe the patient demographic and prescription characteristics. Demographic information included age (< 40, 40–64, ≥ 65 years old), sex, and region of residence; socio-economic status (SES) based on neighbourhood income quintile (from Statistics Canada census 2012 public use files); and types of cancer. Prescription characteristics included prescription medication payer (Pharmacare, Personal Care Home, Palliative Care Home, Home Cancer Drug Program, Employment and Income Assistance, Other (which includes federal public drug benefit programs, Veterans Affairs Canada, out-of-pocket or cash, and private insurance)). SAS software for Windows, version 9.4, was used for all analyses.

Results

Over the entire study period from 2003/04 to 2015/16, a total of 22,393 OAA users with cancer filled 416,382 prescriptions. The characteristics of OAA users with cancer are presented in Table 3. This population was older than the overall Manitoba population (57.1% versus 47.9% over the age of 65 years) and 59.3% were female (versus 64.5% in the overall Manitoba population). When examined by medication category, hormonal therapies were used by more women (64.8%) than men and targeted medications were used by more men (57.5%) than women.

Table 3.

Demographics for all individuals using oral anticancer agents with cancer (2003/04–2015/16)

Demographic	All OAA (N = 22,393) (%)	Traditional (N = 6043) (%)	Hormonal (N = 16,094) (%)	Targeted (N = 1587) (%)	Supportive (HCDP only)* (N = 9431) (%)
Age (years)
≤ 39	4.07	9.22	2.22	6.11	6.2
40–64	38.79	39.35	38.52	48.77	47.04
≥ 65	57.13	51.43	59.26	45.12	46.76
Female sex	59.35	51.20	64.78	42.53	56.71
Winnipeg Regional Health Authority	58.8	58.33	58.98	60.11	59.46
Socio-economic status
Q1 (lowest)	17.7	17.99	17.41	17.96	15.08
Q2	21.9	21.6	22.05	20.16	19.47
Q3	21.58	20.82	21.87	20.86	22.86
Q4	18.23	18.68	18.14	19.28	20.39
Q5 (highest)	19.15	19.87	18.98	20.73	21.67
Percent of prescriptions by payer
Pharmacare	57.47	59.23	58.92	34.68	0
HCDP	27.95	26.14	26.52	50.62	100
Nursing home	6.41	3.76	7.46	0.36	0
Employment/income assistance	2.47	2.91	2.42	1.86	0
Palliative care	1.31	1.78	1.24	1.01	0
Other**	4.38	6.17	3.43	11.46	0

*This category includes supportive care agents and specific anticancer medications covered by the HCDP program from 2012/13 onwards

**This category includes federal public drug benefit programs with separate formularies and coverage such as First Nations and Inuit Non-Insured Health Benefits, Veterans Affairs, and out of pocket or cash prescriptions (these prescriptions may have private insurance coverage) and prescriptions with private insurance

Reference: Raymond C, Leong C, Fransoo R, et al. Outpatient Oral Anticancer Agents in Manitoba. Winnipeg, MB. Manitoba Centre for Health Policy, Summer 2018

There has been a 1.5-fold increase in the outpatient OAA users with cancer from 222.3 per 100,000 to 328.9 users per 100,000 over the study period (Fig. 1). The total number of prescriptions for OAA that were filled increased from 21,895 prescriptions in 2003/04 to 37,878 prescriptions in 2015/16. The majority of OAA users were comprised of users of hormonal medications for the treatment of estrogen receptor positive breast cancer, which increased from 153.8 per 100,000 to 231.1 per 100,000. We observed a decrease in the use of traditional OAAs from 37.1 per 100,000 to 32.7 per 100,000 over the study period and a marked increase in targeted therapies from 2.5 per 100,000 to 33.6 per 100,000 over the same time period (Fig. 2). More specifically, a marked decline in the use of alkylating agents (a traditional OAA) occurred over time from 24.5 per 100,000 to 9.4 per 100,000, while a 10.7-fold increase in the use of protein kinase inhibitors (a targeted OAA) was observed from 2.5 per 100,000 in the first quarter of 2003/04 to 26.8 per 100,000 in the last quarter of 2015/16 (Fig. 3). The trends of specific OAAs used over time are shown in Supplementary Figures 1-7.

Fig. 1.

Crude quarterly rate of oral anticancer agent use among people with cancer (2003/04–2015/16) (per 100,000 people)

Fig. 2.

Quarterly rate of all categories for oral anticancer agents used among people with cancer (2003/04–2015/16) (per 100,000 people)

Fig. 3.

Crude quarterly rate of oral anticancer agent drug use among people with cancer by drug category (2003/04–2015/16) (per 100,000 people)

Most prescriptions were paid for by Pharmacare until the introduction of HCDP (between 80% and 90%). Prior to the introduction of the HCDP program, we observed large variation in rates of users of OAAs in the first quarter of the year as compared with the third or fourth quarter of the year. This pattern was the same for the number of prescriptions. After the introduction of the HCDP, this variation in rates of quarterly users and prescriptions per 100,000 population disappeared. This difference was found to be statistically significant (p = 0.003 for users and p ≤ 0.0001 for prescriptions). When compared to the time period before the HCDP was introduced, the relative difference between the prevalence of OAA users and prescriptions from the highest quarter of the year to the lowest quarter of the year was significantly greater than after the HCDP was introduced (relative rate = 1.05 with upper and lower bounds 1.02–1.08, alpha 0.05, standard error 0.01, chi-squared value 13.37, p = 0.003 for OAA users and relative rate = 1.16 with upper and lower bounds 1.14–1.18, standard error 0.01, chi-squared value 251.81, p ≤ 0.0001 for OAA prescriptions). This suggests that the pattern of OAA user and prescription patterns changed after the HCDP was introduced.

The total expenditures on OAAs increased almost 4-fold from 6.8 million in 2003/04 to 26.1 million in 2015/16. The average annual cost per user increased from $1651 to $4163, and the number of users increased from 4141 (21,895 prescriptions) to 6281 (37,878 prescriptions) from 2003/04 to 2015/16 (Table 4). Targeted therapies were the most expensive agents per user compared with the other categories (Fig. 4). The mean annual cost per user of oral targeted therapies remained constant at approximately $30,000 per user, while the total cost of oral targeted therapies increased from $1,759,957 to $18,929,741 from 2003/04 to 2015/16 (Table 5).

Table 4.

Total cost of oral anticancer agents used by people with cancer

Fiscal year	Total cost	Mean annual cost per user	Number of users	Number of prescriptions
2003/04	$6,837,337	$1651	4141	21,895
2004/05	$8,498,974	$1939	4383	23,891
2005/06	$9,222,118	$2076	4443	26,355
2006/07	$9,588,356	$2090	4587	27,398
2007/08	$11,010,159	$2298	4792	30,368
2008/09	$13,066,054	$2675	4884	32,838
2009/10	$14,842,558	$2903	5113	34,007
2010/11	$16,811,833	$3148	5340	35,838
2011/12	$19,151,161	$3536	5416	36,099
2012/13	$20,996,830	$3669	5723	36,577
2013/14	$22,354,141	$3837	5826	35,818
2014/15	$24,109,216	$3892	6195	37,420
2015/16	$26,146,644	$4163	6281	37,878
2003/04–2015/16	$202,635,383	$3019	22,393	416,382

Reference: Raymond C, Leong C, Fransoo R, et al. Outpatient Oral Anticancer Agents in Manitoba. Winnipeg, MB. Manitoba Centre for Health Policy, Summer 2018

Fig. 4.

Cost per user among people with cancer (2003/04–2015/16)

Table 5.

Total cost of targeted oral anticancer agents used by people with cancer

Fiscal year	Total cost	Mean annual cost per user	Number of users	Number of prescriptions
2003/04	$1,759,957.02	$32,591.80	54	376
2004/05	$2,406,002.13	$38,806.49	62	502
2005/06	$2,462,788.03	$35,692.58	69	513
2006/07	$3,529,871.58	$33,617.82	105	772
2007/08	$4,462,021.20	$31,871.58	140	1001
2008/09	$5,884,723.64	$28,428.62	207	1343
2009/10	$7,272,033.68	$30,049.73	242	1630
2010/11	$9,374,657.87	$31,671.14	296	2095
2011/12	$12,334,885.67	$36,820.55	335	2428
2012/13	$14,070,335.54	$33,106.67	425	2767
2013/14	$16,820,947	$33,441.25	503	3673
2014/15	$16,295,541.71	$29,203.48	558	4041
2015/16	$18,929,741.17	$30,384.82	623	4662
2003/04–2015/16	$115,603,506.87	$31,943.49	1587	25,803

Reference: Raymond C, Leong C, Fransoo R, et al. Outpatient Oral Anticancer Agents in Manitoba. Winnipeg, MB. Manitoba Centre for Health Policy, Summer 2018

The rates of ambulatory physician visits and inpatient hospitalizations were the highest for those receiving traditional OAAs followed by targeted and hormonal agents (Fig. 5 and Fig. 6). A slight decrease in hospitalizations was seen in those taking hormonal OAAs from 30 per 100 people in 2003/04 to 19 per 100 people in 2015/16. The rate of inpatient hospitalizations for those receiving traditional OAAs remained stable from 2003/04 to 2015/16, while the rate of hospitalizations for those receiving targeted OAAs fluctuated between 30 per 100 people and 48 per 100 people during the same time period.

Fig. 5.

Crude ambulatory visit rate for OAA users with cancer diagnosis by drug category (2003/04–2015/16)

Fig. 6.

Crude hospitalizations for OAA users with cancer diagnosis by drug category (2003/04–2015/16). Reference: Raymond C, Leong C, Fransoo R, et al. Outpatient Oral Anticancer Agents in Manitoba. Winnipeg, MB. Manitoba Centre for Health Policy, Summer 2018

Discussion

We found an increase in the use of OAAs, which was also observed in previous research (Timmers et al., 2012, Benjamin et al., 2015). More specifically, we observed a marked increase in the use and total expenditure of oral targeted therapies in this 13-year observation study. We also found a plateau in traditional OAAs and an increase in hormonal therapies for breast and prostate cancer. These trends may reflect changes in diagnosis patterns and cancer incidence over time, as well as an increase in availability of OAAs and a shift towards the use of OAAs for treating certain types of cancers.

A total of 190,847 prescriptions were covered by HCDP since its implementation in 2012. This allowed cancer patients to receive OAAs and eligible supportive therapies without having to pay an income-based deductible every year, which can range from 3.09% to 6.98% of the adjusted total family income (Manitoba Pharmacare Program, 2018). As a result, patients who have met their income-based deductible for the year would often try to have the remainder of their eligible prescriptions filled prior to the start of the new fiscal year (which starts April 1 of every year). This results in a seasonal prescription filling pattern where higher rates of prescription fills are often seen in the quarter prior to April 1. The new policy did not appear to impact the upward linear trend in the use of OAAs; however, the policy changed how and when patients filled these prescriptions. More specifically, the seasonal prescription filling pattern for OAAs was eliminated after HCDP was implemented. Although we did not evaluate adherence with prescriptions, as it is known that adherence with OAAs is high, the change in prescription filling patterns suggests that having to pay out of pocket impacted when patients filled prescriptions for OAAs. For some types of cancers, the oral formulation of hormonal or chemotherapy may be the only option. The HCDP now allows patients to receive these therapies at no cost, just as it would be if they had an intravenous chemotherapy option.

The rising total expenditure on OAAs, and in particular the targeted OAAs, is an important observation from a drug policy perspective. Others have also reported on the dramatic rise in the cost of therapies for cancer over the past decade, largely due to the availability and uptake of new targeted therapies (Dusetzina, 2016, Pataky et al., 2018). However, coverage of these agents varies across the country (Sorin et al., 2019, Sultan, 2019). Like Manitobans, residents living in British Columbia, Alberta, and Saskatchewan also receive coverage for both injectable and oral cancer therapies. Ontario and the Atlantic provinces have variable coverage on orally administered cancer medications (Morgan et al., 2017). National pharmacare strategies have largely aimed at ensuring essential medications are accessible to Canadians but a strategy for addressing the rising cost of high-priced medications has yet to be determined (Morgan et al., 2015). The Patented Medicine Prices Review Board (PMPRB) have also proposed changes for setting the price of medications effective July 2020, including incorporating measures related to financial impact on consumers in determining whether a drug cost is too high, and revising the list of countries used to compare drug prices to reflect countries similar to Canada (Government of Canada, 2020). With the growth of oral agents available in clinical practice, our findings underline the need for addressing these high-cost medications in future developments of a national drug program. It is not known whether private insurance companies will continue to cover the cost of expensive medications in the future and what impact this might have on programs such as HCDP.

We observed higher rates of ambulatory visits and hospitalizations for traditional OAA users with cancer compared with targeted and hormonal OAA users. There is no indication of how often the hospitalizations were related to the OAAs or their side effects.

This study has several strengths including a population-based linked administrative dataset that allows for research evaluating the impact of health policy on the pattern of prescription utilization and cost. The administrative databases used in this study were also not restricted to age or income assistance, which can be a limitation in other jurisdictions conducting similar research. However, similar to other population-based studies, there were some limitations to consider. All data included in this report are derived from contact with the healthcare system. Because not everybody who seeks medical attention and receives a prescription for a medication actually fills the prescription, this may underestimate the number of prescriptions written for OAAs and specific supportive therapies in Manitoba. Medication use not captured in the DPIN system may include physician samples, although the possibility is low for these types of medications. There are some systems of compassionate use or dispensing outside of DPIN system for OAAs; notable examples include thalidomide which is dispensed through the pharmaceutical company and oral fludarabine which is not in DPIN as it is dispensed through the CancerCare Manitoba pharmacy. Some individuals are not included in the DPIN system, including, for example, those who are incarcerated or members of the RCMP or those who do not register for the provincial Pharmacare program. These individuals make up a very small portion of the Manitoba population. It should be noted that some Personal Care Homes (approximately 25%) do not fill prescriptions at community pharmacies (and are therefore not included in the DPIN system). DPIN also does not distinguish out-of-pocket expenses by payer (e.g., deductibles paid by the patient versus expenses paid by private insurance) and it was not possible to specifically estimate deductibles paid by HCDP after its introduction. As a result, it was difficult to estimate the added expenses incurred by the province post-HCDP implementation (which does not cover expenses paid by private insurance).

For several analyses, we considered medical diagnoses through administrative data. The use of administrative data may underestimate the prevalence of a given condition in the population, because these definitions require a patient to seek contact with the healthcare system. There is the potential for underestimating the prevalence of disease in a population as physician visits could at the time of the study only result in a single billing code. There is also a potential for underestimating or overestimating a given condition due to misclassification. We were not able to explore the impact of this change on OAA adherence, including primary non-adherence where patients have a prescription written but never filled, due to data limitations. It would be of interest to examine the cost trade-off between OAAs and other cancer treatments in the future, including intravenous chemotherapy (and the cost associated with administration of these agents). We were limited in our ability to study this given that the data on these variables to make accurate cost comparisons between cancer treatment modalities are unavailable.

Conclusion

In this large-scale population-based analysis, we saw an increase in the use of and expenditure on OAAs, in particular, the targeted therapies. While the HCDP program saved Manitobans thousands of dollars in out-of-pocket expenses on medications for cancer through eliminating the cost of the deductible to patients, and in some cases, the entire cost of the prescription, our findings underline the need for addressing these high-cost medications in future developments of a national drug program.

Funding

Funding for this work was received from the Manitoba Health, Seniors and Active Living (MHSAL).

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Disclaimer

The results and conclusions are those of the authors, and no official endorsement by the MCHP, MHSAL, or other data providers is intended or should be inferred.