The use of willingness to pay to assess public preferences towards the fortification of foodstuffs with folic acid

Abstract

Objectives To assess public attitudes towards the fortification of flour with folic acid, and quantify their intensity of preference towards the proposed policy.

Design Structured interviews describing the proposed policy of fortification followed by questions about the respondent's preferences towards fortification.

Setting and participants  A United Kingdom community sample of 76 people interviewed at home.

Main variables studied Direction of preference towards fortification, willingness to pay (WTP) for the preferred course of action, and the reasons behind their preferences.

Results  Responses showed that 51 (67%) were in favour of fortification, 15 (20%) were opposed, while the remaining 10 (13%) were either indifferent or were unsure. Those in favour of fortification tended to be younger and poorer than those opposed to it. Willingness to pay estimates show that those in favour of food fortification had more intense preferences, with mean and median WTP around twice as great. Reasons for being willing to pay were centred on the health benefits, with particular reference being made to the intervention saving lives and it being preventative. Those opposed tended to believe that there was insufficient evidence.

Conclusions The combination of a policy vote, WTP and qualitative data, allow us to assess the direction, intensity and motivations behind people's preferences. Further work needs to be undertaken to gather more robust estimates of public preferences for fortification, and to better understand attitudes towards public health interventions more generally.

Introduction

Evidence demonstrating that increased use of folic acid preconceptionally and in the first few weeks of pregnancy can reduce the incidence of neural tube defects (NTDs) has been available for many years. ¹ Such evidence formed the basis of the current recommendation for women who could become pregnant to take 400 μg/day folic acid as a supplement. This evidence and the possibility of fortifying flour with folic acid, as a public health response, was recently reviewed in the UK by the Committee on Medical Aspects of Food and Nutrition Policy (COMA). ² Key conclusions of the report were:

• • There is a link between folate status and the risk of NTDs.
• • Increased folic acid intake may delay diagnosis of undiagnosed vitamin B₁₂ deficiency, which could hasten neurological damage associated with this condition. The neurological damage could range from subacute combined degeneration (SCD) and peripheral neuropathy through to spinal cord damage and severe disability.
• • The universal fortification of flour at 240 µg/100 g in food products as consumed would reduce the incidence of NTD‐affected pregnancies by 41%, without resulting in unacceptably high intakes in any group of the population.

The proposal to fortify wheat flour with folic acid raises many issues, such as freedom of choice and the trade‐off between the benefits to newborn babies and the risks to people suffering from undiagnosed vitamin B₁₂ deficiency. Consequently, the Department of Health undertook a consultation exercise focussing on the proposal to fortify wheat flour with folic acid. ³ , ⁴ The exercise was based around 12 questions and covered many issues such as acceptability, technical feasibility, cost, consumer choice and legislative feasibility. There were 167 respondents from a range of academics, consumer groups, trade organizations, individual companies, health authorities and individuals, although the general public were not the main focus of attention.

Of the 142 respondents to the question asking about the desirability of fortification, 59% agreed with the proposal, 30% disagreed and 11% were unsure or had reservations. Some variations in preferences across the sample were noted, such as consumer groups and individuals being less inclined to be in favour of the proposals, whilst academics and professional bodies were more inclined to support them. The main reasons given for supporting the proposals were their impact on unplanned pregnancies and the benefits to women. Concerns were raised about the effect on vitamin B₁₂ deficiency, the impact of higher intakes on children and lack of consumer choice.

When evaluating the desirability of any health intervention, economists balance the benefits of the intervention with the costs of implementing it; this process is a key part of economic evaluation. Such evaluations typically confine their investigation to health benefits. For example, a cost‐effectiveness analysis of the fortification of flour may assess cost per NTD‐affected birth avoided, whilst a cost‐utility analysis may assess the cost per quality adjusted life year gained of the child and/or parents. Using this information, a decision‐maker could ascertain how to maximize health benefits for a given budget (which economists refer to as ‘technical efficiency’). However, such evaluations fail to assess non‐health benefits or whether the additional benefits are worthwhile purchasing (this latter issue is referred to as ‘allocative efficiency’ by economists). The importance of non‐health benefits is investigated by this study, as too is the possibility of assessing whether fortification is a worthwhile undertaking.

This study builds on the consultation exercise by developing that work in two ways. First, preferences are measured and valued using the willingness to pay (WTP) technique. ⁵ This approach is being increasingly used in health economics because of its ability to measure more than just health‐related preferences, for example, the value of information, ⁶ , ⁷ altruism and choice. ⁸ It also allows intensity of preference to be measured, rather than just the direction of preference which the consultation exercise expressed through statements of support or opposition for fortification. This is important when considering this intervention which raises issues of health benefit, risk of harm, equity, altruism and freedom of choice. Secondly, our study is based on a community sample, as opposed to the purposive sample of the consultation exercise.

Methods

The WTP data were collected as part of a short interview conducted on the respondent's doorstep. A postal questionnaire was not considered appropriate in this study as the description of fortification is complex, and would be more likely to produce low response rates and misunderstandings if presented in that way. A community sample was used, and although it was not a true probabilistic random sample we aimed to interview people from a variety of social groupings. The sample was drawn from three electoral wards within Sheffield that included low, middle and high income areas. Households were selected by sampling at random from a list of streets within the chosen wards, and then the houses were sampled by the interviewer when visiting the selected streets. All interviews were undertaken between 9.00 a.m. and 6.00 p.m. No sample size was pre‐specified, instead, the maximum number of interviews that could be completed over a 4‐week period were undertaken.

The interview was designed to take no more than 20 min to complete. It began with general introductory information followed by the presentation of the fortification scenario (see Appendix 1). Respondents were encouraged to read the scenario themselves. The scenario had to omit many details of the potential effects of folic acid fortification and the possible ways in which it could be implemented, in order to avoid cognitive overload. Consequently, the scenario did not mention the possible reduction in the risks of cardiovascular disease and neuropsychiatric conditions, the possibility of non‐compulsory fortification and the option of combined supplementation with vitamin B₁₂. ²

This was followed by a ‘policy vote’ question asking whether the respondent supported or was opposed to fortification. Those respondents who were in favour were then asked if they would be willing to contribute anything in extra taxation for fortification going ahead. Taxation was used as the payment vehicle in order to make the questionnaire as plausible/realistic to respondents as possible. It was felt that in view of the mass public health intervention nature of fortifying flour with folic acid that the costs of doing so would not be passed onto consumers in the form of higher retail prices (thereby risking lower levels of consumption), but rather the costs would be borne by the government out of taxation revenues.

If respondents were not willing to pay they were asked to state why. Those who indicated a WTP were then asked to state the maximum amount they would be willing to contribute each year for fortification to proceed. Respondents were asked to indicate their maximum WTP value on a payment card. This card replicated the scenario already shown to the respondent, together with a linear scale ranging from zero to £100. An option of stating an amount in excess of £100 was also available on the payment card. The choice of the zero to £100 scale was based on previous work on a similar topic. ⁸ This method of eliciting a respondent's maximum WTP was chosen over the alternatives of open‐ended questions and dichotomous choice questions, as the former has been shown to produce erratic valuations, ⁹ whilst the latter requires much larger sample sizes than that used in this study. ¹⁰ Finally, respondents were asked to explain why they were willing to contribute.

If respondents indicated that they were not in favour of fortification, they were asked if they would be willing to pay additional food prices for non‐fortified foods to be made available. The choice of private expenditure as the payment vehicle here (as opposed to taxation) was again down to a desire to make the questionnaire as realistic as possible to respondents. It was felt that asking people to consider paying extra taxes to ensure availability of non‐fortified products was less intuitively appealing than asking them to consider paying higher retail prices (consumers are familiar with paying higher prices for goods which are different from the norm, e.g. organic produce). If respondents were not willing to pay, they were asked to state why. Those who indicated a WTP were then asked to state the maximum amount they would be willing to contribute each year for fortification to go ahead. As with those who supported fortification, WTP was indicated on a payment card.

Reasons for not being willing to pay were independently classified by both authors as either ‘protest’, ‘zeroes’ or ‘unsure’. Protesters refuse to give a value for the benefits of fortification, and were assigned as such if they did not make any reference to the benefits of the fortification in their reply. Implied zero values were assigned to individuals if they indicated a consideration of the benefits of fortification, but did not, or could not, pay. Individuals who were undecided about whether they were willing to pay or not for the benefits of fortification were classed as being unsure.

The interview was completed by asking respondents how they rated the difficulty of the WTP procedure, together with sociodemographic questions. Finally, respondents were offered further information on folic acid, produced by the Health Education Authority, ¹¹ to allay any concerns which may have been raised by the interview.

Results

Seventy‐six people were interviewed from 165 households that answered the door, giving a response rate of 46.1%. A further 304 households were approached but did not answer the door. The sociodemographics of the respondents are shown in Table 1. The overall sample contains large proportions of individuals who have degree‐level education and with relatively high annual incomes (i.e. >£25 700).

Table 1.

Sociodemographics by response group

	In favour n = 51 (%)	Opposed n = 15 (%)	Indifferent or unsure n = 10 (%)	All n = 76 (%)
Gender
Male	27.5	26.7	30.0	27.6
Female	72.5	73.3	70.0	72.4
Age (years)
16–35	39.2	0.0	30.0	31.1
36–55	27.5	69.2	50.0	37.9
56+	33.3	30.8	20.0	31.1
Education
‘A’ level and    higher	58.0	66.6	70.0	61.3
Below ‘A’ level	42.0	33.4	30.0	38.7
Income
Less than £6800	23.3	23.1	20.0	22.7
£6800–£11 100	16.3	0.0	20.0	13.6
£11 100–£17 100	18.6	30.8	10.0	19.7
£17 100–£25 700	18.6	23.1	10.0	18.2
More than    £25 700	23.3	23.1	40.0	25.8
Employment status
Paid employment	34.0	40.0	50.0	37.3
Housework	16.0	13.3	10.0	14.7
Unemployed	2.0	0.0	20.0	4.0
Student	18.0	13.3	0.0	14.7
Retired/pensioner	30.0	33.3	20.0	29.3
Number of dependent children
0	66.7	66.7	70.0	67.1
1	15.7	0.0	20.0	13.2
2	11.8	20.0	0.0	11.8
3	5.9	6.7	10.0	6.6
4	0.0	6.7	0.0	1.3

Policy vote question

Responses to the policy vote question showed that 51 (67%) were in favour of fortification, 15 (20%) were opposed, while the remaining 10 (13%) were either indifferent or were unsure. Those in favour of fortification tended to be younger and poorer than those opposed to it.

For those respondents who were in favour of fortification, 26 (51%) were willing to pay to ensure the programme went ahead, 8 (16%) implied a zero valuation, 8 (16%) protested to paying, whilst the remaining 9 (17%) were unsure about paying. The protests were in the form of people failing to accept the choice and saying that it is a government issue, or there are other ways of tackling and funding the problem (e.g. ‘Manufacturers should pay or simply increase the price of food’ and ‘There are other ways of doing it’). The comments implying zero valuations were varied and included statements that it is not very important (e.g. ‘There are other issues that I feel more strongly about’), and that the respondent was too poor to pay any money. Two respondents raised the issue that they themselves were not going to benefit from the intervention (e.g. ‘It would only benefit pregnant women and not all of society’ and ‘Because I don’t need it personally').

Reasons for being willing to pay for fortification mainly centred on the health benefits or more general notions of it ‘being a good thing’, with particular reference being made to the intervention saving lives and it being preventive. Other respondents raised issues of ‘social responsibility’, the relative importance of interventions for children, and the need to reduce the burden on the NHS.

For those respondents who were opposed to fortification, 6 (40%) were willing to pay extra for non‐fortified foods, 2 (13%) indicated zero valuations, while the remaining 7 (47%) respondents either protested or were unsure about paying. Included among the protests were two respondents who indicated the need for individual responsibility (e.g. ‘People should know about folic acid already’ and ‘People already have good access to information, it is generally available’).

Reasons for being willing to pay for non‐fortified foods included the belief that it was not necessary (e.g. ‘It is not necessary at the current time’), that there was insufficient evidence (e.g. ‘I would want to know the outcome of further research’ and ‘It can’t be proven. I have a general problem with science, it is all lies!') and a distrust of additives (e.g. ‘I am against universal additives in principle’).

Willingness to pay estimates

A crude comparison of WTP estimates shows that those in favour of food fortification had more intense preferences, with mean and median WTP around twice as great (Table 2). In addition to assessing the relative intensity of preferences, WTP data can be used when assessing the overall change in welfare brought about by the intervention. However, such an evaluation would require additional information on the costs of the intervention, averted costs to the health service and changes in consumption patterns, and is beyond the scope of this paper.

Table 2.

Summary willingness to pay results

	n	Mean (SD)	Median	Sum
In favour	32*	22.8 (26.3)**	12.5	730.0
Opposed	8*	11.9 (16.8)**	6.3	95.0

^* Excludes protests. **t‐tests, P = 0.272.

It is widely recognized that WTP figures tend to give additional weight to the preferences of the rich, because of the fact that WTP is positively associated with ability to pay. Whilst most WTP studies have not addressed this issue, ¹² an approach has been developed whereby a set of income‐based weights can be estimated. ¹³ This approach is not adopted for this study because its size precludes the use of Donaldson's approach. However, as it is those who are opposed to fortification who tend to have higher incomes, a re‐weighting would reduce the value of their preferences further, thus increasing the measure of net benefit.

Explanatory analysis was also undertaken to investigate differences in the WTP responses between population groups. Because of the small sample sizes regression analysis was limited to the WTP responses of those who were in favour of fortification (Table 3). No respondent characteristics were statistically significant, although the direction of effects suggested that income was found to be positively related to WTP, as was being female and having higher educational achievement.

Table 3.

Explanatory analysis of willingness to pay responses of those in favour of fortification

Parameter	n	Parameter estimates	95% Confidence intervals	P‐value
Intercept	29	58.9	19.6, 98.2	<0.01
Age (years)
16–35	12	−13.8	−52.2, 24.6	0.46
36–55	7	3.1	−32.7, 39.0	0.86
56+	10	0	–	–
Gender
Male	9	−21.6	−49.5, 6.3	0.12
Female	20	0	–
Educational level
Below ‘A’ level	11	−15.4	−43.6, 12.8	0.27
‘A’ level and above	18	0	–
Children
No children under    16 years	19	−9.1	−37.1, 19.0	0.51
Children under    16 years	10	0	–
Income
Less than £6800	8	−16.4	−55.5, 22.7	0.39
£6800–£11 100	5	11.8	−30.3, 53.8	0.57
£11 101–£17 100	5	−22.7	−61.1, 15.7	0.23
£17 101–£25 700	5	−28.5	−65.9, 8.8	0.13
More than £25 700	6	0	–	–

Difficulties identified by respondents

For those in favour of fortification, 7 (29.2%) found it difficult answering the WTP question, while of the remainder 17 (70.8%) did not find the question difficult. Similar proportions were found in those opposed to fortification (33.3% vs. 66.7%), although only six respondents were available. Reasons for finding it difficult included not knowing what the additional tax would be used for (e.g. ‘I would not know exactly where my money was going, and who else would be paying’) and general disagreement with the concept (e.g. ‘It is a daft question, a measure of strength from very strong to indifferent would be better than using monetary values’). Despite this, several respondents commented on the interview being undertaken well (e.g. ‘The interview was done in the right manner’ and ‘The interview was conducted in a reasonable manner’).

Other comments

When asked for any other comments about the interview, several people expressed an interest in the topic (e.g. ‘I found it very interesting, I will now read up on folic acid’), hoped for action (e.g. ‘I hope that something is done’) or more information (e.g. ‘How would the total cost of adding folic acid to flour compare with the total cost of providing better education for women?’). Another theme that arose was the participants' view of being involved in the decision‐making process, with polarized views being expressed (e.g. ‘It is good to question people, I am concerned about such issues’ and ‘It is good to see this issue being brought to light. The more we know, the better’, in contrast to, ‘I will say again, it is an NHS issue’).

Discussion

This study represents the first attempt to capture public preferences on the proposed fortification of wheat flour with folic acid in order to reduce the incidence of NTDs. In addition to a straightforward policy vote question, WTP and qualitative information were also collected. However, these results can only be considered as hypothesis generating, due primarily to the sample size and frame. In addition to the sample having a disproportionately large number of households with high incomes, the use of unsolicited doorstep interviews may have contributed to a non‐response bias. While this does not necessarily invalidate the methodological findings of this study, any further study that wishes to inform policy must use a larger and more representative sample.

Another problem with this approach is the need to simplify the information given in the fortification scenario. The selection and presentation of information in the scenario presented to respondents was fairly arbitrary, and it could be that presenting different information could produce significantly different responses. For example, we could have given a more pessimistic description of the potential effects of undiagnosed vitamin B₁₂ deficiency, or described all the hazards and benefits that may be associated with fortification. However, including all information is more likely to confuse respondents, and could trigger further negative responses because of the lack of clear‐cut evidence on several issues relating to fortification. The potential biases produced by the omissions from, and the language of, the scenario need to be carefully considered in future work. Conveying health risks to the public is just one area where recent experience has shown the sensitivity of public sentiment to seemingly innocuous phraseology. ¹⁴ Piloting work is essential in order to identify which facets of the intervention are deemed important and how these should be described.

Another source of bias which may have been present is that arising from the use of two different payment vehicles to elicit WTP values, i.e. taxation for those in favour of fortification and higher retail prices for those opposed. For example, the use of taxation as a payment vehicle may introduce bias because of the fact not everyone will pay the same amount of taxation. Those with low incomes and a low tax burden may bid more in recognition that the valuation exercise is hypothetical and they will not actually have to pay. While recognizing the potential for such bias, this should be set against the biases which may have occurred had respondents felt the WTP questions were not plausible/realistic. If the questions had been regarded as unrealistic, it is more likely that fewer people would have responded, and that those people who did respond would have given less thought to their values, thereby raising doubts over whether the WTP amounts represent an accurate reflection of preferences. The relative impacts of these sources of bias is an empirical question – one which cannot be answered using data in this study.

In defence of the decision to use different payment vehicles, only one person objected to the method of payment; a respondent in favour of fortification thought that food prices should be increased instead of using taxation. Furthermore, there were no differences in the proportions of respondents confronted with each payment vehicle in terms of difficulty in answering the WTP questions.

A more fundamental problem with the responses given is that they may not represent ‘true’ preferences. Watson and colleagues interviewed women on the issue of folates in pregnancy and hypothesized that an adaptive process towards the assimilation of information was present. ¹⁵ An initial ‘emotional response’ to sensitive health information was proposed that changed over time once the women had chance to better understand the risks and benefits presented to them. If the notion of an adaptive process is accepted, then the validity of the responses given within interviews of this type is cast into doubt. However, the explanatory analysis lends some support to the validity of the responses by showing the responses to be in line with the prior expectation that richer respondents were willing to pay more. Such a finding is frequently used as a weak test of validity in WTP studies. ⁵

Despite this, the study demonstrates the feasibility, and desirability of gathering these data when assessing public health interventions of this type. The combination of a policy vote question, WTP data and qualitative data allow us to assess the direction, intensity and motivations behind people's preferences. Whilst there are problems with this approach, it represents an important step forward in an area of health research where the input of consumers is notoriously sparse. ¹⁶

These data could be used for several purposes. First, the policy vote question clearly and unambiguously shows the level of support for fortification, and when combined with respondents' comments regarding the need for further information, demonstrate public education needs.

Secondly, the WTP data could in principle be used in a future cost‐benefit analysis of fortification. While there remain a number of issues to be resolved with WTP, ¹⁷ valuing the preferences of all members of the public is essential for public health interventions of this type. ¹⁸ Cost‐effectiveness and cost‐utility analyses would effectively exclude the preferences of those opposed to fortification and those not accruing any health benefits. Given the size of opposition to fortification, and conversely, the expressions of support from those not directly affected by the intervention, including the preferences of these groups is very important.

Thirdly, this study, together with previous work looking at fluoridation, ⁸ has demonstrated our ignorance about the public's perception of public health. Recurrent themes in the two studies are individuals' sense of altruism, their aversion to additives and restriction of freedom, their inclination to give added weight to the health benefits of children, and many other fundamental issues. More work should be undertaken to develop a public health strategy that is more in line with the views of the society which we represent.

It is interesting to note that the WTP estimates shown in Fig. 1 are quite different from those produced by a similar study of community water fluoridation. ⁸ The fluoridation study demonstrated that the minority possessed much stronger preferences than the majority, who supported fluoridation. In this study, intensity of preference was greatest in the majority group, and as such, the WTP and policy vote data reinforce each other. Such a result provides a less ambiguous expression of support for fortification than in the fluoridation study, and clearly demonstrates the potential value of gathering WTP information data on public health interventions of this kind.

Figure 1.

Willingness to pay responses, excluding protests.

One other difference between these two studies is the number of respondents who protested against the WTP question by refusing to give a WTP. Respondents in the fluoridation study were more likely to protest towards paying than those in this study (31% vs. 20%). Whilst this may reflect local disenchantment with the local water authority at the time of the fluoridation study, it may also indicate a greater willingness to engage with the topic of folic acid fortification. Further work is required to assess the motivations behind these ‘protesters’. However, care must be taken when classifying respondents as ‘protesters’. Although our classification was based on two independent assessments, it is possible that other researchers may interpret the responses in different ways. It would also be preferable for any future work looking at individual's motivations and values to use more valid qualitative research techniques.

This independent study was designed partly in response to the proposed format of the Department of Health's consultation exercise on folic acid, which did not systematically approach a community sample. The consultation exercise was targeted at industry and other interest groups, which meant that of the 167 responses made to the exercise, only 29 (17%) were from individual members of the public. ⁴ We feel that data of the kind produced by our study could add valuable information to a consultation exercise, for example, the proportion of the population that support an initiative, the intensity of their support and, importantly, the reasons why people oppose it. Such data could also be used more formally in any cost‐benefit calculation undertaken by the Department of Health when evaluating the policy.^*

Future consultation exercises on a wide range of issues should consider the use of WTP information in the format used here, although further refinements in the methods need to be incorporated. With respect to folic acid, further work needs to be undertaken to gather more robust estimates of public preferences for fortification, and more clearly assess the public's personal value systems and their views on public health interventions more generally.

Appendix 1. Description of fortification of food with folic acid and its effects

Every year around 180 babies in the United Kingdom are born with neural tube defects. There are two main types of neural tube defect – anencephaly and spina bifida. The majority of babies born with anencephaly die within a few days of birth. Those babies with spina bifida are born with a range of disabilities ranging from mild disability to those which are severely disabled. Furthermore, the presence of these abnormalities will cause many miscarriages, and many other pregnancies will be terminated following ultrasound scans diagnosing neural tube defects.

Adding folic acid to the diet of women prior to conception and for the first few weeks of pregnancy can reduce the risk of neural tube defects. All women should be advised to take folic acid supplements prior to conception. However, not all women are given this advice or follow it.

An alternative approach is to add folic acid to food. If folic acid is added to food, it has been estimated that the number of babies being born with neural tube defects can be reduced by around 74 every year. The number of miscarriages and terminations would also be expected to fall.

If this goes ahead, it is possible that the diagnosis of one disease – vitamin B₁₂ deficiency – in the elderly may be delayed because of the higher levels of folic acid masking the disease. This could lead to people with the disease experiencing some loss of sensation in the arms and legs, but many doctors feel that this is avoidable.

If fortification goes ahead, the folic acid will be added to flour, and so only products which contain flour will contain the additional folic acid. Some flour supplies will not be fortified, and products which contain non‐fortified flour will be clearly marked on their packaging. The taste and appearance of food will not be altered in any way by the addition of folic acid.