Measurement of consumer preference for treatments used to induce labour: a willingness‐to‐pay approach

Abstract

Aim

The purpose of the study was to assess the acceptability to consumers of two methods of induction of labour using a willingness‐to‐pay (WTP) approach. The methods compared were amniotomy plus oxytocin and prostaglandin E₂ vaginal gel, followed by oxytocin if necessary.

Methods

A description of each method was presented, in questionnaire format, to pregnant women attending a public hospital ante‐natal clinic. Women were asked to choose one of the two treatments, then give a valuation in dollar terms for both their preferred treatment and the alternative.

Results

It was found that 73.7% of patients preferred gel. The mean maximum WTP for amniotomy plus oxytocin was Aus$133 while that for gel was Aus$178 (P=0.0001). Those who chose amniotomy plus oxytocin were WTP 90% more for their preferred treatment compared with the alternative (Aus$180 vs. Aus$95). Similarly, those who preferred gel were WTP 90% more for their preferred treatment compared with the alternative (Aus$222 vs. Aus$119).

Conclusion

Consumers were able to assess drug information provided on the two therapies, make an informed choice and to value that choice. Information obtained in this way, combined with information on costs, could be used in policy decision‐making.

Introduction

Recent research, in the UK, has recognised the importance of incorporating the consumer perspective into health‐care decision‐making at a policy level. Opinion polls, patient satisfaction surveys and the measurement of consumer preference have all been attempted, the last approach providing the most suitable information to aid policy decision making. ¹

Consumer preference associated with health‐care interventions has usually been measured in terms of patient utility (satisfaction) using techniques such as the standard gamble and time trade off. ² Quality of Life Year values (QALY’s) have usually been elicited. ³ However, these techniques are not appropriate or applicable in all circumstances. ^{1 , 4 , 5} Willingness‐to‐pay techniques, however, offer advantages in that they are based on economic theory ⁶ , 7, ^–8 and can incorporate process issues as well as health outcomes. ⁹

Because the economic market does not function well in publicly funded health‐care systems, the Contingent Valuation (CV) approach has been used to elicit a monetary value of such health care programmes. ⁸ CV involves presenting respondents with hypothetical but realistic scenarios about the interventions under investigation and asking them to give a valuation in monetary terms, as though such a market does exist. The valuation is expressed as the maximum willingness‐to‐pay (WTP) for those interventions. ^{8 , 10} This approach has been used, to some extent, to elicit values for interventions involving pharmaceuticals. ¹¹ , 12, 13, 14, 15, 16, ^–17

PGE₂ vaginal gel is used to ripen (soften and dilate) the cervix prior to the induction of labour and may be enough on its own to induce labour, or may have to be followed by oxytocin. Prior to the introduction of the gel in Australia in 1991, the standard treatment for induction of labour was amniotomy followed by oxytocin. The acquisition cost of the gel is 15 times that of oxytocin.

Some evidence of improved health outcomes following the use of prostaglandin F₂α and PGE₂ has emerged, however, this evidence has been based on a meta‐analysis in which the results of controlled trials of all prostaglandins, given by any route, were combined. ¹⁸ Other benefits to patients, which may justify the increase in cost incurred if the gel is used, have been suggested by the manufacturer of the gel. ¹⁹ Some of these benefits could be assessed as health outcomes, e.g. the proportion of patients having a Caesarean section, but others are more related to the process undergone by the patient. For instance, freedom of movement during labour may contribute more to patient satisfaction than whether a Caesarean section was performed or not.

The specific aim of this study, therefore, was to use a contingent valuation technique to measure both the direction and strength of preference, in dollar terms, for either amniotomy plus oxytocin or prostaglandin E₂ gel, in a way which took both process of care and health outcome into account.

Methodology

Ethics approval for the study was obtained from the Ethics Committee of Central Sydney Health Service. Information from the Obstetric Information System was obtained on disk for all women (approximately 3000) who underwent induction of labour in King George V Hospital, a Sydney teaching Hospital, in the 4 years from 1990 to 1993 inclusive.

Qualitative work – the description of treatments

Twelve of these women who had had at least one of the methods of induction in the past were interviewed about their experience. The women chosen were mixed in their demographics and health outcomes so that there was a wide range of experience and opinion.

The interviews took place in women’s homes and were audio‐taped at the time. They were later transcribed and content analysed for themes. Of particular interest were the advantages and disadvantages of each method of induction because, from these, the possible trade‐offs women faced could be identified. From a combination of information in the literature and the content of these interviews, a description of the two treatment approaches was developed. This description, expressed in terms of six attributes, was then reviewed by the Medical Director of Obstetrics and Gynaecology at King George V Hospital and, after some minor adjustments, was incorporated into the survey instrument to provide information upon which women could base their choice of treatment (Appendix 1 1).

Table 1.

Appendix 1  Description of treatments

The study population consisted of pregnant women attending the public ante‐natal clinics of King George V Hospital. At the time the study took place, neither the researcher nor the women knew which patients would require an induction. Women were approached by the researcher in the clinic and asked to participate in the study. If interested, they were given a questionnaire to complete while waiting to see the obstetrician or midwife.

The survey instrument

The questionnaire was piloted on 40 pregnant women attending the antenatal clinic to see if participants could:

(a) understand the information given to them;

(b) make a choice of treatment;

(c) give a valuation in dollar terms expressed as their willingness‐to‐pay for their preferred treatment and the alternative.

In addition, the pilot study was used to estimate the sample size.

This pilot questionnaire was analysed and, as a result, adjusted in very minor ways to enhance the clarity of the instructions to respondents. Based on the numbers of responses required for a subsequent section of the questionnaire (not reported here) a sample size of 300 was determined. (This was based on a power of 80% and significance level of 0.05%.) Taking into account a response rate of 75%, as assessed in the pilot study, the final sample size was calculated to be 400.

Choice and willingness‐to‐pay (WTP) questions

Respondents were asked to read an introductory explanation, then a fairly detailed description of the two alternative approaches to the induction of labour (Appendix 1 1). They were then asked to choose between the treatments.

Next they were asked to indicate the amount of money they were sure they would be willing to pay for their preferred treatment, and how much for the alternative, by marking the amount with a tick on a categorical scale ranging from Aus$0 to Aus$1500 (Appendix  2). This variant of the payment card method was chosen because alternatives have either been found too difficult for people to manage successfully (open ended question) ^{4 ,   20} or to require too large a sample size (take‐it‐or‐leave‐it approach). ²¹ The range of values used in this study was established as a result of qualitative work, in which women were asked to state how much they would be willing to pay for each treatment. If the value was greater than Aus$1500, they were asked to state the actual amount they would be willing to pay. The scale used approximated a log scale. Respondents were asked to value both treatments, not just the one they preferred, because previous studies, which asked for a valuation of only the treatment given, have failed to discriminate between close alternatives. ⁴

Table 2.

Appendix 2  Payment vehicle. Please place a tick (✓) against the amount you are sure you would be willing to pay for the treatment you prefer and for the other treatment

Respondents were not told the actual cost of the treatments but were told that the researchers were interested in their valuation in dollar terms. Although in the community setting, in Australia, consumers are required to pay an out of pocket cost for pharmaceuticals, whether subsidised or not, in Australian hospitals they are not required to pay directly. They only pay a Medicare levy, as part of the taxation system. Because of the concern of the Ethics Committee, respondents in this study could not be asked how much they were willing to pay directly out of their own pocket in case they thought that they really did have to pay. Instead, they were asked to imagine that they were in a country like the USA, where they did have to pay directly out of their own pocket.

Patient demographics

Demographic details were requested as shown in Table 1. Respondents were asked their combined annual household income, before tax, as one of 10 categories and their level of education as one of seven categories. These seven categories were reduced to two categories for Table 1. Relevant information about their childbirth history was also elicited (Table 1).

Table 1.

 Patient demographics

Importance of aspects of treatments

Having completed the valuation task, respondents were asked to indicate the importance to them of various aspects of the treatments, as described in Table 2, on a four point Likert scale ranging from ‘not important’ at all to ‘very important’.

Table 2.

 Importance of aspects of treatment

Feedback about the questionnaire, views on natural childbirth and choice

Finally, their reaction to the questionnaire itself was elicited. Respondents were asked to indicate whether they agreed or disagreed with certain statements about the information given on a four point Likert scale ranging from ‘strongly agree’ to ‘strongly disagree’. These statements asked if women found the information easy to understand, if they were pleased to receive the information and if they found the questionnaire too long to complete. In addition, they were asked if they would like to have a natural childbirth if possible and if they would like to have a choice about the treatment they would receive. The four categories were then reduced to two, ‘agree’ and ‘disagree’, and the proportion of respondents with each view was calculated.

At the end of the questionnaire women were invited to make any comments they wished about the questionnaire itself, or any issues it raised for them. These comments were analysed for content and themes.

Data analysis

The response rate to the questionnaire and the proportion choosing each method of induction were calculated. The mean maximum WTP for each method was calculated. The maximum WTP was taken to be the mid‐point between the category chosen and the next highest category. This more accurately reflects maximum WTP and also makes it possible to treat the values as point estimates rather than as categorical data, an approach which has been adopted by other workers in the field. ⁹

The WTP values were then log transformed to normalise the data. Normality of the distribution was determined using the Lilliefors test (a modification of the Kolmogorov–Smirnov test). Visual confirmation of normality was obtained by examination of the normal Q‐Q plot, detrended normal Q‐Q plot of the transformed data. A paired Student’s t‐test and the corresponding 95% confidence intervals (CI) of the difference in means were computed on the log transformed data to compare the overall WTP for treatments A and B. One‐way analysis of variance was performed on the difference between the log WTP for treatment A and that for treatment B in order to test whether this difference was dependent on the respondent’s preference group. The mean difference of the log WTP between the two methods was then estimated for each preference group and the ratio of WTP for preferred treatment/alternative was calculated.

A backward stepwise ordinary least squares (OLS) regression analysis on the log WTP values for each treatment was used to identify predictors of these values. Predictor variables were excluded from the model if the level of significance of the T statistic exceeded 0.05. Possible predictors fell into four categories: those related to the demographics of respondents such as age, education, income, a language other than English spoken at home; those related to childbirth history such as stage of pregnancy, previous children, induction or Caesarean section; those related to the importance of various aspects of the treatments; and those related to the attitude of respondents to the information provided, to natural childbirth and to the choice of treatment in the future. Given the statistical properties of a log‐normal distribution, the exponential of the log WTP equation at the mean values gives the median of the distribution of WTP and not the mean. To estimate the mean of the untransformed dependent variable, WTP, the median was weighted (multiplied) by exp(σ²/2), where σ² is the variance of the model. ⁹

All the analyses were carried out using SPSS software Version 7.5.

Validity

The theoretical validity of the WTP question, that is, the extent to which it was consistent with economic theory, was tested. First, consideration was given to whether the variables entered in the regression analysis influenced WTP as might be expected. Secondly, the effect of income on Willingness‐to‐pay for each treatment was investigated. A priori hypotheses tested were that:

(a) the more important women thought it was to avoid having their membranes ruptured, the greater their WTP for gel;

(b) the more important women thought it was for women to avoid a labour longer than 8 hours, the less their WTP for gel.

Results

The overall response rate was 359/400, i.e. 89% of questionnaires given out. However, not all responses were considered to be valid. Only those who had made a choice of treatment and given a valuation for the preferred treatment and for the alternative were considered valid. Furthermore, the valuation for the preferred treatment was required to be at least equal to or more than the valuation of the alternative. Some respondents did not make a choice of treatment (18). Other respondents valued the alternative more highly (56) or did not fill in all parts (58) and it was felt that they had not completed the task successfully (Table 3). They were, therefore, omitted from the subsequent analysis, taking the total number of respondents to 243, i.e. 68% of questionnaires returned.

Table 3.

 Number of completed responses categorised according to tasks completed

Patient demographics

Respondents ranged in age from 15 to 44 years with a median age of 29 years (Table 1). 72/241 (29.9%) spoke a language other than English at home. Between them these women spoke 32 different languages. 124/241 (51.7%) had a post‐school educational qualification. The modal category of household income was Aus$31 000–Aus$40 000. The median stage of pregnancy was 26 weeks with a range of 6 weeks to 42 weeks. 102/239 (42.7%) women were pregnant with their second or subsequent child, 49/238 (20.6%) had undergone induction previously and 14/238 (5.9%) had had a previous Caesarean section (Table 2).

Importance of aspects of treatment

This is shown in Table 2, along with the mean score for each aspect.

Feedback about questionnaire, views on natural childbirth and choice

Over 90% of women were pleased to receive the information contained in the questionnaire and found it easy to understand and complete. Only 13% thought that it took too long to do. 91.2% said they would prefer to have a natural childbirth if possible but interestingly 96.6% of respondents said they would like to have a choice of treatment if they had to undergo induction of labour in the future.

28.8% (70/243) of respondents took the time and trouble to make extra comments at the end of the questionnaire. Of these, 8 referred to the WTP question or to cost. Two of these comments were explicit objections to ‘user pays’ type questions within the public health‐care system and two were comments about the difficulty of placing a monetary value on pain or the safety of the unborn child, etc. The remainder related to doubts or confusion respondents had about their own responses.

Choice and willingness to pay (WTP)

Of the 243 women 179 (73.7%) chose treatment B (gel) while 64 (26.3%) chose treatment A (amniotomy + oxytocin).

The mean maximum WTP overall for treatment B (gel) was Aus$177.72 (CI 95% Aus$153.21, Aus$206.15) whereas that for treatment A was Aus$132.74 (CI 95% Aus$113.73, Aus$154.93) (P=0.000) (Table 4). Since log transformation of the data resulted in its normalisation, as indicated by the normal Q‐Q plot and detrended normal Q‐Q plot, statistical tests were conducted on the log transformed data.

Table 4.

 Maximum willingness‐to‐pay values for all respondents and for different preference groups

The analysis of the 243 responses included approximately 5% of patients (13) in each group who gave a valuation of Aus$0 for both treatments. Two of these 13 women made an extra comment at the end of the questionnaire relating to the WTP question. It is debatable whether these 13 should be considered as zero values, or protests to the technique. When this group of 13 was excluded from the analysis, 170/231 (73.9%) of those remaining chose gel and (60/230) 26.1% chose amniotomy plus oxytocin. The mean WTP for gel was Aus$202.59 and that for amniotomy plus oxytocin was Aus$148.84 (P=0.000).

When preference for treatment was taken into account, the ratio of mean WTP for preferred treatment over the alternative was similar for the two preference groups (1.9) (Table 4) and indicated that women were willing to pay 90% more for their preferred treatment, regardless of treatment preference.

Influences on WTP – ordinary least squares regression analysis (OLS)

For the valid responses with a full data set, three variables relating to the demographics, aspects of treatment or views, when entered in the regression model, were found to have a significant influence on the log WTP for gel (Table 5). These variables were income, the importance of avoiding having the membranes ruptured artificially, and avoiding a labour longer than 8 hours. The adjusted R² for the model was 0.14. The mean WTP for a woman with the mean income, mean score for the importance of avoiding artificial rupture of the membranes, and mean score for the importance of avoiding a long labour was Aus$239.06.

Table 5.

 Influence of aspects of treatment and patient demographics on Log WTP for Treatment B (Gel) − Ordinary Least Squares Regression

Only one variable had a significant influence on Log WTP for treatment A (amniotomy plus oxytocin). This was income. The adjusted R² for this model was 0.08.

Validity

The results were consistent with economic theory. First, the variables entered into the regression analysis influenced WTP as expected. For example, the more important it was to women to avoid having their membranes ruptured, the greater their WTP for Treatment B (gel) (P < 0.05). Similarly, the more important it was to them to avoid a labour longer than 8 hours, the less the WTP for Treatment B (P < 0.05). These results were consistent with the information given to women about the two treatments.

Secondly, from the OLS regression analysis it was clear that income was positively correlated with WTP for both treatments.

Discussion

In this sample of pregnant women, the majority (73.7%) of patients chose gel. The mean maximum WTP for gel was Aus$178, compared with Aus$133 for ARM plus oxytocin (P=0.0001). In addition, both preference groups were, on average, willing to pay 90% more for their preferred method.

WTP in health

Several objections to the use of WTP have been raised in the literature. One objection is that it is difficult for respondents to place a monetary value on life, and pain and suffering, etc. ³ A second objection is that health‐care systems in most countries are publicly funded and there is concern that respondents may assume that a system whereby the user pays will be introduced if WTP questions are asked. This can lead to protest votes, or zero valuations, even from hypothetical statements. ⁷

In this study 5% (13/243) of responses analysed gave a zero valuation for both treatments. When the results were reanalysed without these responses included, the mean WTP values for both treatments, as expected, increased (Aus$149 vs. Aus$133 for treatment A and Aus$178 vs. Aus$203 for treatment B) although the proportion choosing each treatment remained the same. Whether these responses should be included or not is unclear. The fact that these women valued both treatments at zero may indicate that they did not wish to engage in the WTP process and, therefore, were expressing a protest vote. However, only two of these 13 women gave extra comments at the end of the questionnaire which indicated an objection to the WTP question being asked within a publicly funded health‐care system and, thus, could clearly be identified as protest votes.

In addition, the fact that women were asked to ‘imagine’ that they had to pay for the treatments out of their own pocket, rather than actually pay for them, may have influenced their WTP values. If the respondent thinks that the provision of the product depends on the valuation of the product but does not depend on the level of personal payment, valuations higher than the true maximum WTP may be given. ⁷ It may be interesting to compare the values obtained in this study with values elicited from women attending a private hospital in Australia, where the treatment may be paid for directly.

Furthermore, this method may have given women licence to express values inconsistent with their income. Despite the possible lack of realism, a positive correlation was found in this study between income and WTP for the preferred treatment.

A third objection to the use of WTP methods in health is that the approach most suitable for the valuation of health‐care interventions and, more specifically, those involving pharmaceuticals has not been agreed. A recent review of contingent valuation studies in health‐care found that, of the 42 studies identified which used a WTP question, 19% used the payment card technique, 38% an open‐ended question, 26% the take‐it‐or leave‐it (closed‐ended) approach and 29% bidding games. ²² WTP has been used widely in environmental and transport studies in cost benefit analyses and many different valuation elicitation techniques have been tried. ²³ , 24, 25, 26, ^–27 Guidelines on the design and implementation of contingent valuation studies, produced in 1993 by the National Oceanic Atmospheric Administration, recommended that valuations should be elicited using a take‐it‐or‐leave‐it approach and that in‐person interviews rather than mail surveys be used. ²⁰ However, it is not always feasible to follow both these recommendations at the same time, since the take‐it‐or‐leave‐it approach requires a large sample size ²¹ and in‐person interviews are time consuming and expensive to conduct.

All four main valuation elicitation methods have their advantages and disadvantages. Bidding games have been criticised because it is thought that the starting bid may influence subsequent bids. ¹¹ For a similar reason, the payment card technique, and its variation as used in this study, has been criticised because it is thought that the range of amounts presented may influence the response. ⁷ However, it has been argued that providing respondents with a range of values reflects the everyday life situation more accurately, since consumers would usually decide their WTP for a commodity from a range of known prices. ⁴ When an open‐ended question has been asked respondents have often relied upon their perception of the price of the product being valued when stating their WTP and this may be different from their maximum WTP. ⁴ In addition to the range, the actual values presented in the payment scale technique may influence the response.

Does this method discriminate between close substitutes?

The approach adopted here was based on the novel technique used by Donaldson and co‐workers ⁴ in their estimation of WTP for carrier screening for cystic fibrosis. Although Donaldson and co‐workers used an open‐ended question, both options under consideration were valued by the same individual. This approach was used by Donaldson because previous studies associated with clinical trials, in which patients valued only the care they received, did not discriminate between interventions. ^{10 , 12 , 27} This was thought to be because patients were being asked to value their own care compared with no treatment, rather than compared with the alternative treatment. ⁴

Although no significant difference was found by Donaldson and co‐workers (1997) in the overall WTP for the screening methods evaluated, the workers considered the method to be more discriminating, in that extra information was obtained about preferences and about the WTP of each preference group. They found that the majority group was WTP 68% more for their preferred method while the minority group was prepared to pay 90% more for their preferred method and concluded that this information could conceivably be useful in policy decision‐making.

In the current study, the majority preferred gel and the WTP for gel was significantly higher than the WTP for the standard treatment. Thus, the method did discriminate between treatments. In addition, both the majority and minority groups expressed a willingness‐to‐pay 90% more for their chosen treatment than for the alternative.

However, it should be remembered that 56/357 respondents (15.7%) gave a valuation for their preferred treatment which was less than their valuation for the alternative and were excluded from the analysis. For Donaldson and co‐workers (1997) 23% of responses fell into this category.

Should these responses be considered rational and included in the analysis or should they be excluded? The reason for such responses in the current study is not known. It could be that these valuations reflected respondents’ true feelings, or could have been due to their perception of cost of the two treatments as suggested by Donaldson and co‐workers (1997). These workers found that, particularly with an open‐ended question, some respondents tended to estimate the cost of the good being valued. In the current study, a specific question about difficulty with cost was not asked.

Other reasons could be that the layout of the questionnaire was confusing or, alternatively, that Treatment B is not always expected to work on its own while Treatment A is. One woman indicated verbally that she valued Treatment A more highly because it could be relied upon, even though she would like to try Treatment B first. In future studies it may be advisable to ask respondents to comment on their reasons for their preference structure, including specific questions on the influence of the perception of cost on their valuation.

Other researchers have since modified this method. Donaldson and co‐workers ²⁸ have asked women to value their preferred type of maternity care, rather than their less preferred type, while Davey and co‐workers ¹⁷ (1998) asked respondents to place a value on an additional user charge for their preferred type of insulin. This may have made it easier for respondents to complete the WTP task since only a choice and one valuation was requested. The results from the current study show that 94% of respondents completed the choice and valuations of preferred treatment successfully while only 68% successfully completed the valuation of the alternative as well as the other two tasks.

The value of information

The information provided to patients could influence their response and may be part of what is being valued by respondents. Therefore, it is important that the information itself does not introduce bias. Also, a balance must be struck between sufficient detail and information overload. While it is acknowledged that outcomes for women who have had vaginal births previously, or for those with a dilated cervix, may be better than for those having their first child, or those with a closed cervix, no explicit distinction in outcomes was made in this study. This more global, average approach was adopted here because the description, obtained from patients who had received the treatments in the past and reviewed by an expert obstetrician, was already quite complex and detailed.

Use in policy decision‐making

How could such a result be applied in policy decision‐making? Should PGE₂ vaginal gel be made available to consumers attending this hospital? Although this sample of women was typical of the English speaking inner city population of Sydney, it is not known how representative of the whole maternity hospital population, or wider population the women were, and thus, some doubt exists as to whether the results are ‘generalisable’. Unfortunately non‐English speaking women could not be included in the study because translation into too many different languages (>20) would have been required. These results may, therefore, only be valid for the English speaking population of the maternity hospital.

Cost benefit analysis attempts to inform a social valuation question and the aim is to ensure a socially efficient allocation of resources. ²⁹ The normative welfare criterion, on which cost benefit analysis is based, is the Potential Pareto Improvement criterion. ^{3 , 6 , 29} A strict application of this criterion would result in the treatment chosen by the majority, in this study Treatment B (gel), being provided. Those who preferred Treatment B could compensate for those who preferred Treatment A and, in theory, a net gain could be incurred. ⁴

However, when both direction of preference and strength of preference data are available the size of the majority, the relative strengths of the preferences and the added cost of meeting the preferences of both groups (assuming the treatments are equally effective) could also be taken into account. In this study, the majority preferred gel and the WTP for gel was significantly higher (1.3 times higher) than that for the standard treatment, amniotomy plus oxytocin. Thus, a net gain of 30% could be achieved.

When costs are considered, it is difficult to see how a net gain overall could be incurred, even if the only difference in cost involved were in the acquisition cost, because the acquisition cost of gel is 15 times that of oxytocin. If a decision to provide gel were made, on the basis that the majority preferred it and that its average cost per patient is relatively small (approximately Aus Aus$70), because gel cannot be expected to always induce labour on its own, the standard treatment would also need to be provided. The standard treatment would, therefore, be available for the minority, who prefer it to gel, and provided for them because the cost would be less than if gel were used.

Although this methodology is in a developmental stage and several issues remain unresolved it has the potential to enable researchers, health‐care decision‐makers, practitioners and manufacturers to gain some insight into consumer preferences for particular drug treatments. As a consequence it may be possible to incorporate these preferences into health care decision‐making.

Acknowledgements

We are very grateful to Dr Andrew Child, the Director of Obstetrics and Gynaecology in King George V Hospital for his support for this project. In addition we would like to thank Mai Lee, the Nursing Unit Manager of the Outpatient Clinics, and her staff for their co‐operation during the data collection period in the clinics. Finally, we would like to thank the Pharmacy Board of NSW for their financial assistance.

Appendix 1

Description of treatments

1 [Table]

Appendix 2

Payment vehicle

Please place a tick (✓) against the amount you are sure you would be willing to pay for the treatment you prefer and for the other treatment

2 [Table]