Contingent valuation of inhaled insulin: a Canadian perspective

Summary

The objective of this study was to assess the preference and willingness-to-pay (WTP) for inhaled insulin from a random sample of general public perspective in Ontario, Canada. This was carried out using a mail survey using the contingent valuation method. Significantly more respondents preferred inhaled insulin (n=114) over subcutaneous insulin (n=6; p<0.001). The mean monthly WTP for inhaled insulin was CAN$68.59 ± 44.65 (95% confidence interval CAN$58.87–78.07), significantly more than the average subcutaneous insulin cost in Ontario of CAN$50. The WTP for inhaled insulin in the diabetic subgroup was CAN$98.52 ± 48.57, which is significantly higher than that of the general public (p<0.001). Multiple regression analysis showed a strong association between respondents’ income and diagnosis of diabetes and their WTP for inhaled insulin (p<0.001). Major influencing factors were convenience and household budget. The results of this study suggest that the general public in Ontario prefer inhaled insulin and are willing to pay significantly more per month than the current cost of subcutaneous insulin.

Keywords: :

• contingent valuation
• willingness to pay
• diabetes
• inhaled insulin

Introduction

Diabetes is a substantial burden on patients, society and healthcare systems1. In Canada, the burden of diabetes was estimated to be approximately CAN$13 billion in 1998, mainly due to long-term complications2. The goal of therapy for diabetes is to control blood glucose levels to near normal values and to reduce the risk of vascular complications3. The importance of intensive therapy in the prevention of diabetes-related complications has been established4,5. All type 1 diabetic patients require exogenous insulin. In addition, as the disease progresses, most patients with type 2 diabetes require insulin to control their blood glucose levels.

Insulin as a peptide cannot be used orally6. Hence, many innovations have been developed to overcome such a problem. Although better blood glucose control has been achieved with insulin analogues, multiple injections of intensive treatment are inconvenient. Alternative routes of insulin administration have been studied, including nasal and transdermal7–13, but inhaled insulin has been the most promising of all alternative routes14, as it can pass easily from the lung alveoli into the blood15. The efficacy and safety of inhaled insulin in controlling blood glucose levels have been established in clinical trials, and one powder form of inhaled insulin has been approved for patient use by the US Food and Drug Administration and European Agency for the Evaluation of Medicinal Products16–19.

At the time of this study, inhaled insulin was not available in Canada. Similar to many new drugs, it was expected that inhaled insulin would be more expensive than insulin injection because of the higher doses of insulin combined with the higher cost of the delivery system. However, studies have shown that both type 1 and type 2 diabetic patients are more satisfied with inhaled insulin compared with injected insulin20–22. This suggests that, despite the higher price, some diabetic patients might choose inhaled insulin.

Diabetic patients’ willingness to pay (WTP) for inhaled insulin has previously been reported using the contingent valuation method23. However, owing to the fact that the Canadian healthcare system is a Beveridge system, Canadian guidelines for economic evaluation of health technology recommend the use of the general public perspective in pharmacoeconomic studies24. The objective of this study was to determine the preference and WTP of the Ontario general public for inhaled insulin.

Methods

A contingent valuation survey was conducted to assess the value that the Ontario general public assigns to inhaled insulin. An ex-ante (non-user-based) perspective was adopted and the comparator was rapid-acting insulin analogue (RAA) injection. The study protocol and instruments received the approval of an independent Ethics Review Board.

Study design

Between November 2004 and February 2005 an invitation letter was mailed to 250 rural and urban households in Ontario, Canada, randomly chosen based on postal codes. In the invitation letter, residents were invited to participate in a mail survey on diabetes drug therapy. Residents who were interested in participating in the survey were asked to sign a consent form and to reply to the investigator. Upon receiving written consent, the survey package was mailed to the participants. Inclusion criteria were:

• • being an Ontario resident;

• • having the ability to read and understand English;

• • being >18 years of age; and

• • having the ability to provide informed consent.

Survey instrument

The survey instrument was developed based on a previously tested self-administered questionnaire used in the assessment of diabetic patients’ WTP for inhaled insulin23. The survey instrument was developed in two phases. A clinical expert reviewed the instrument for content validity and the instrument was pilot tested with 25 Ontario residents. The survey instrument consisted of the following sections: (1) background information about diabetes and its treatment; (2) the choice scenario; (3) attributes of the comparator products; (4) WTP and debriefing questions; (5) the payment scale; and (6) a demographic and clinical questionnaire. Standardised information on diabetes, including management, prevalence in the province of Ontario and insulin therapy, was provided to participants in a three-page illustrative informative letter produced by the investigator.

In contingent valuation studies, a hypothetical market (scenario) where participants can buy the product of interest is described25. In this study, participants were asked to consider themselves in a situation where they were diagnosed with diabetes, were prescribed insulin and were deciding to buy insulin. The participants were told that they had the option of choosing either insulin injection or inhaled insulin. The alternatives were described in terms of the timing of administration, convenience and blood glucose control, attributes that have been used in prior contingent valuation studies of insulin23,26,27. In the scenario, participants were informed that inhaled insulin and subcutaneous insulin had similar efficacy and adverse drug reaction profiles and that they might need a bedtime injection with both regimens (see the Appendix).

Participants were told that payment would be ‘out-of-pocket’. This payment vehicle was chosen because it was expected that inhaled insulin would not be covered by the Ontario Drug Benefit Plan (ODB) and consequently by some third-party payers28. The concept of out-of-pocket payment for pharmaceutical products is familiar to Ontarians. Approximately 10–12% of Ontarians do not have a drug plan and many third-party drug plans require patients to pay the drug cost and get reimbursed later. Participants were asked to consider a monthly payment for their preferred product. The participant's economic limitations and the chronic nature of diabetes were highlighted to avoid an unrealistic response.

In WTP studies, the offering of a single point estimate to participants and asking them whether they are willing to pay that amount for their product of choice can lead to estimate bias29. To minimise the starting point bias and also to facilitate the participants’ decision-making process, the ‘payment scale’ method was used to elicit participants’ maximum WTP for their product of choice30. In this method, a chart containing a wide array of potential WTP amounts ranging from zero to very large amounts is used. Payment levels for this study were determined in the author's previous study. To avoid ‘range bias’, bids were vertically arranged with the lowest bid at the top and the last bid was an open-ended question for values above CAN$700. The bid values increased with variable constant rates (higher increasing rates with bigger dollar values). The bids ranged from CAN$0 to CAN$700 (year 2005 dollars), with an additional response option for values more than CAN$700.

To evaluate the reasoning behind their responses to the WTP questions, respondents were asked the following debriefing questions:

• •What was important to you when you chose your drug of choice?

• •What was important to you when deciding the amount that you were willing to pay for your drug of choice?

Statistical analysis

Data from the questionnaire were entered into a spreadsheet by the investigator and double checked by a peer. Descriptive statistics were calculated for socioeconomic and clinical data and the WTP estimates. Qualitative responses to the debriefing questions were categorised for analysis by the investigator.

The proportion of participants preferring inhaled insulin was assessed by a one-sample χ2 test. The mean WTP was tested using a single-sample t-test in which the mean was compared with the average monthly cost of insulin injection (CAN$50). Various analyses of participant subgroups were undertaken. Differences in preference for inhaled insulin between the general population and participants with diabetes and other chronic diseases were assessed using χ2 tests. Differences in maximum WTP between the general population and participants with diabetes and other chronic diseases were assessed with the Mann-Whitney U-test. The association between maximum WTP and respondents’ socioeconomic and clinical characteristics was assessed using multiple regression analysis. The independent variables in the analysis were set a priori as income (considered as a continuous variable measured at ratio level); diagnosis with diabetes; experience with self-injection; and experience with inhaler. A type I error level of 5% was chosen for statistical significance. SPSS statistical software (SPSS Inc., Chicago, IL) version 12 was used for all statistical analyses.

Results

Of the 150 (60%) questionnaires returned, 120 (80%) were complete and used for analysis. Of the 30 questionnaires that were eliminated from the analysis, 12 had incomplete sociodemographic information making it impossible to characterise the respondent and 18 had no response to the WTP questions. The latter group expressed their ideas against raising cost of pharmaceutical products in general.

The mean age (± standard deviation) of participants was 54.8 ± 12.2 years and 53% were male. Approximately 79% of participants had post-secondary education and 72.5% had a full-time job. Approximately 69% were married or in a common law relationship. Approximately 86% of participants had a drug coverage plan (health insurance covering drugs) either by private insurance or the ODB. Approximately 72% of participants had an annual household income above CAN$60,000. Participants estimated their mean monthly out-of-pocket payment for prescription drugs to be CAN$34.41 ± 13.39. These findings are presented in greater detail in Table 1.

Table 1. Socioeconomic characteristics of survey participants (n=120).

Characteristic	n (%)
Mean (± sd) age (years)	54.8 ± 12.2
Gender
Male	64 (53.3)
Female	56 (46.7)
Marital status
Single	11 (9.2)
Married/common law	83 (69.2)
Separated/widowed	26 (21.7)
Education
High school or less	25 (20.8)
College/university	95 (79.2)
Employment
Full time	87 (72.5)
Part time	21 (17.5)
Unemployed/retired	12 (10)
Annual household income (CAN$)
0–20,000	5 (4.2)
21,000–40,000	4 (3.3)
41,000–60,000	25 (20.8)
61,000–80,000	34 (28.3)
81,000–100,000	37 (30.8)
>100,000	15 (12.5)
Drug plan^*
Yes	103 (85.8)
No	17 (14.2)

sd, standard deviation; CAN$, Canadian dollars. ^* Plan (insurance) for drug cost coverage.

A total of 31 participants (25.8%) had diabetes, of which 19 (15.8%) participants had experience with self-injection, all of whom were using insulin pen injection. More participants had experience with drug inhalers (21.6%). These findings are presented in more detail in Table 2.

Table 2. Health-related characteristics of survey participants (n=120).

Characteristic	n (%)
Experience with inhalers
Yes	26 (21.7)
No	94 (78.3)
Experience with self-injection
Yes	19 (15.8)
Use of insulin	19 (100)^*
No	101 (84.2)
Chronic disease
Yes	37 (30.8)
No	52 (43.3)
Diabetes	31 (25.8)

^* As a proportion of participants with self-injection experience.

Of the 120 participants 114 (95%) preferred inhaled insulin over insulin injection. This proportion was significantly greater than 50% of random choice (p<0.001). The mean amount that participants were willing to pay each month for inhaled insulin was CAN$68.59 ± 44.65 (95% confidence interval CAN$58.87–78.07). This was CAN$18.59 more than the typical monthly cost of RAAs (CAN$50, based on the average cost of RAA pen cartridges in three retail pharmacy chains in Ontario and the dose of insulin for average diabetic patients) and this difference was statistically significant (t_df95=3.83). Subgroups analysis showed that all participants with diabetes preferred inhaled insulin compared with insulin injection. The mean WTP for inhaled insulin in the subgroup of participants with diabetes (CAN$98.52 ± 48.57) was significantly more than the mean WTP for inhaled insulin in the general public (Mann-Whitney U=365.0; Z=–4.0; p<0.001).

Responses to the debriefing questions are summarised in Table 3. The majority of participants who preferred inhaled insulin over insulin injection stated that it was either because of its convenience (71.1%) or to avoid the pain of injection (25.4%). The major influence on the amount that participants were willing to pay was household budget (77.2%). Participants who preferred injection over inhalation as the route of administration for insulin (n=6) did not put any value on inhaled insulin. Two participants mentioned that the drug should be covered by a drug plan, and four participants stated household budget limitation for not using inhaled insulin.

Table 3. Analysis of responses to the debriefing questions.

Debriefing question	Preference	Response	n (%)
What was important to you in choosing the type of insulin delivery system that you chose?	Inhaler (n=114)	Convenience of inhaled insulin	81 (71.1)
		Avoiding pain of injection	29 (25.4)
		Needle phobia	4 (3.5)
	Injection (n=6)	Comfortable with injection	6 (100)
What was important to you when deciding the amount that you were willing to pay for the type of insulin delivery system that you chose?	Inhaler (n=114)	Household budget	88 (77.2)
		Health benefits gained	26 (22.8)
	Injection (n=6)	Should be covered through insurance premium or tax	2 (33.3)
		Household budget	4 (66.7)

In the regression model, the multiple correlation coefficient (r=0.806) was significantly different from zero: F=24.153; df (6&78); p<0.001. The proportion of variance accounted for by the model (R2) was 0.650. Table 4 summarises the coefficient of the model of which the coefficient for income and diagnosis with diabetes was statistically significant (p<0.001).

Table 4. Parameter estimates for the regression model (n=120).

Variable	Regression coefficient	se	95% CI	p-value
Constant	–103.957	34.600	–172.840, –35.074	0.004
Income	29.791	2.981	23.857, 35.725	0.001
Experience with self-injection	–2.699	10.203	15.085, 37.056	0.792
Experience with inhaler	–12.624	7.559	–27.672, 2.424	0.099
Diabetes	16.730	4.680	7.413, 26.048	0.001

se, standard error; CI, confidence interval.

Discussion

Diabetes is a costly disease with substantial economic impact on healthcare systems owing to its long-term vascular complications. Hence, optimal management of blood glucose levels of diabetic patients and the prevention of diabetic complications has a direct economic benefit for healthcare systems and society.

The majority of participants in this study preferred inhaled insulin over insulin injection (95.0%; 114/120). The results of this study are consistent with the previous finding in a diabetic patient population at a diabetes clinic in Ontario as well as patient satisfaction studies on inhaled insulin, which have shown that 81% of patients with type 1 diabetes and 71% of patients with type 2 diabetes preferred inhaled insulin over insulin injection22,23,31. The main reasons for choosing inhaled insulin were convenience and to avoid the pain of injection. Those who preferred insulin injection stated their comfort with self-injection as the major reason for their preference. Previous studies have shown that diabetic patients have a less than optimum compliance to therapy, especially insulin, mainly due to fear and sometimes inconvenience of injection32,33. Because inhaled insulin is a non-invasive route of administration, it has the potential to increase patient compliance and reduce long-term diabetic vascular complications15. However, long-term clinical trials are needed to support the positive outcomes of inhaled insulin on diabetic vascular complications. Also, the use of inhaled insulin might be the preferred and safe route in subgroups of patients who have difficulty with self-injection, namely patients with impaired vision or manual dexterity and in children34.

The average amount that participants in this study were willing to pay out-of-pocket each month for inhaled insulin was significantly more than the monthly cost for insulin injection in Ontario. However, it was significantly less than diabetic patients’ WTP for inhaled insulin (CAN$153.70 ± 99.90) that was previously reported. One explanation for the lower valuation of inhaled insulin by the general public in Ontario compared with diabetic patients could be the individual's perception of health and familiarity with chronic disease management. Noticeably, the individual's understanding of the complexity of chronic disease management might affect their preferences towards the attributes of pharmaceutical products. Another explanation for this lower valuation could be related to the intrinsic factor related to insulin therapy. In this study, the efficacy and adverse events associated with inhaled insulin were compared in par with insulin injection. In other words, in this study the route of administration was assessed not the drug. When the long-term outcomes of clinical use of inhaled insulin are available, the individual's valuation might be different as one might see a significant difference (if any) between the two administration routes.

Participants’ WTP for inhaled insulin was positively correlated with their household income. This finding is consistent with the statement made by 77% of participants that household budget was a major factor in influencing the amount they were willing to pay. A positive association between income level and WTP has been reported in several other WTP studies on pharmaceutical products, including insulin formulations27,35,36.

Few studies have assessed the WTP for insulin products. WTP for pre-mixed insulin products has been evaluated in two studies, one from Canada and one from Australia26,27. In both studies, diabetic patients preferred the pre-mixed insulin formulation because of its convenience, which was due to reduced daily injections and elimination of the mixing procedure. The Canadian study showed that diabetic patients were willing to pay two times more (CAN$37 vs. CAN$14 in year 2000 $) for the pre-mixed formulations. The present study also showed that diabetic patients prefer a more convenient product and are willing to pay for its convenience.

Several studies have assessed the WTP for pharmaceutical products using the contingent valuation method. A recent review of published contingent valuation studies on pharmaceutical products showed that the majority of published studies have used the patient perspective and a minority have used the general population perspective, and only one study assessed both the patient and general population perspective37. In the only study with both perspectives, Prosser et al evaluated the parents’ (as a proxy of the patient) and the general population's preference and WTP for a pneumococcal vaccine38. In the Prosser study, the WTP for the vaccine was similar between the parents and the general population. However, when the WTP for inhaled insulin was assessed for the author's studies, there was a significant difference between the patients’ and general population's WTP for inhaled insulin. One explanation for the difference between these findings and the Prosser study could be the ‘scope’ of the study or the nature of illness (chronic vs. acute therapy). The perception of one-time payment for vaccination is very different from the lifetime payment for a drug such as insulin to manage a chronic disease like diabetes. Another related explanation could be the difference in the patient population. As highlighted by Prosser, both parents and the general population might have had a ‘parental guilt’ when questioned about the children's well being, whilst this study only included the adult population. Participants’ ‘knowledge’ of the disease could also be an important factor in the thought process. Patients with chronic disease in general, specifically diabetic patients, are usually well informed of their condition and might have a more accurate assessment of the attributes of the products in the study. Further studies comparing the two perspectives can provide a clearer assessment of the differences between public and patient perspective in WTP studies.

Despite all attempts at accuracy, this study has some limitations. This survey was conducted by mail and in the province of Ontario. The participants might not be an accurate representation of the Canadian general population as they had a higher level of education (79.20 vs. 43% college/university) and relatively higher income (72% with income above the national average of $60,500) than the average Canadian based on the Statistics Canada census; therefore, they were likely to be more interested in the advances in drug therapy or had vested interest in chronic disease management. Language and literacy is another limitation of this survey, as the participants were asked to read and understand a text, and also to comprehend the concept that was described in the scenario. The language selection may have imposed a cultural bias in this study, as Ontario is a multicultural province. Furthermore, it has been argued that the mail surveys might impose a sample selection bias39. Similar studies in countries with different healthcare systems might have different results. For example, in countries with limited or no price controls on pharmaceutical products where patients are exposed to high prices, the WTP amount might be significantly higher. In contrast, the results of similar studies in countries with more comprehensive national drug plans might be less than the Canadian results.

In this study, the general concept of ‘inhaled insulin’ was assessed. However, there are different technologies under development and their unique characteristics could influence participants’ responses. Also, the scenario elicited participants’ preference for preliminary use of a new pharmaceutical product, which would not necessarily represent their long-term preference. In addition, the product attributes used in the scenario were based on limited published data. At the time of this study, inhaled insulin products were not available on the Canadian market. Hence, the contingent valuation survey was the most appropriate method for assessing the value of the non-marketed product in this survey (i.e. inhaled insulin). The scenario used in this method describes a contingent market in an attempt to provide a simulation of a market where the study participant is in a decision-making position. The hypothetical nature of the survey might have created some unrealistic responses. Although the survey scenario was designed to be a simulation of a real-life situation, it may not exactly represent patients’ true purchasing behaviour. Finally, payment scales for eliciting WTP amounts are potentially vulnerable to biases relating to the range of the numbers used in the payment scale. Thus, in the future, after inhaled insulin products have been marketed in Canada, it would be worthwhile to conduct a comparative study and to assess the actual purchasing behaviour towards inhaled insulin to evaluate the predictive validity of the contingent valuation method.

The WTP values found in this study could be used as the measure of ‘benefit’ in a cost–benefit study of inhaled insulin compared with insulin injection. A cost–benefit analysis of inhaled insulin could aid drug plan managers in making informed decisions about adding inhaled insulin to their formularies.

Conclusions

The purpose of this study was to determine the value that members of the general public place on inhaled insulin. The study found that members of the general public were willing to pay a substantial amount out-of-pocket to use inhaled insulin, indicating the high value that they place on the convenience and other non-health attributes of an inhaled versus injectable dosage form of insulin. However, the general publics’ WTP for inhaled insulin was significantly less than that of patients with diabetes, showing the importance of ‘individual's knowledge’ in the decision-making process and valuing a pharmaceutical product.

Appendix

The contingent valuation scenario

A new medical inhaler (puffer) is available that diabetic patients who need insulin can use for controlling their blood sugar level. Patients would take insulin from a puffer just before each meal. It will control the patient's blood sugar level similar to the insulin injections. Using the puffer is safe, but more time and experience is needed to be completely assured. Alongside using the puffer, some patients might still need one bedtime insulin injection. The inhaled insulin would not be covered by the drug plans. If patients choose to use this new puffer, they would have to pay for it out-of-pocket. The characteristics of the two products are described below:

Insulin by injection

• 1. You inject insulin 5–10 minutes before each meal.

• 2. You must eat food after taking the drug.

• 3. You need one injection before each meal.

• 4. You have to measure the dose each time, which depends on your blood sugar level and the food that will be consumed.

• 5. You have to carry an insulin pen with cartridge and an alcohol swab, but injection is easy.

• 6. You need a bedtime injection to avoid a low blood sugar level.

Insulin by inhalation

• 1. You inhale insulin 5–10 minutes before each meal.

• 2. You must eat food after taking the drug.

• 3. You need to take multiple puffs from the inhaler before each meal.

• 4. You need to adjust the number of puffs based on your blood sugar level and the food that will be consumed.

• 5. You have to carry a bulky but easy to use inhaler and the insulin units for the inhaler.

• 6. You may need one bedtime injection to avoid a low blood sugar level.

Based on the abovementioned information, should you need insulin, which product would you choose?