Public views on priority setting for High Cost Medications in public hospitals in Australia

Abstract

Objective  To gather information about views of members of the general public about access to High Cost Medications (HCMs) in public hospitals.

Methods  A structured questionnaire was administered to members of the general public. Individuals were approached in train stations, shopping centres and different venues in the Sydney metropolitan area. People were eligible to answer the survey if they were: over 18 years of age, Australian permanent residents and able to complete the questionnaire in English.

Results  Two hundred people completed the survey. Of these 56% were females, 47% were married, 84% spoke English at home, 88% were working either full‐time or part‐time, 61% had a university degree, 27% had a household annual income greater than $100 000 and 68% had private health insurance. Participants considered factors such as treatment outcomes, quality of life and current health status when determining who should have access to HCMs. Participants wanted resources to be allocated to provide the ‘greatest benefit to the greatest number of people’. Almost half the respondents did not want direct involvement in decision‐making, however, 38% did.

Conclusions  The results offered support for indirect involvement through the development of a process to involve community members in discussion on policy on the provision of treatment and services within health‐care institutions and specifically, to seek the views of members of the public on the provision of HCMs and expensive services within public hospitals.

Introduction

Australian public hospitals provide the major proportion of inpatient care for the Australian population. Due to funding arrangements, supply of High Cost Medications (HCMs) in public hospitals may be challenging. The environment is characterized by capped budgets, scarce resources and cost shifting between different funding systems. ¹ , ² , ³ , ⁴ HCMs may be defined in a number of ways. Health‐care decision‐makers in public hospitals in Australia categorized them in two different ways: medications with a high acquisition cost used in low volume and medications with low acquisition cost but used in high volume. ⁵ New South Wales (NSW) hospital decision‐makers perceived that they were primarily concerned about those medications that had an impact on a capped and limited budget. ⁵

According to Peeters there seems to be a lack of community participation in prioritization and rationing in health care in Australia. ⁶ There have been attempts to improve public participation at different decision‐making levels. ⁷ The National Medicines Policy (MP) recognizes the fundamental role of consumers in achieving the objectives of the policy. One of the main four aims of the NMP is access to medicines. ⁸

Initiatives by the Commonwealth Government to engage consumers in heath care (in general) have included: the Consumer and Provider Partnerships in Health Project (CAPPS), the Consumer Participation Conference in May 2001 and the creation of The National Resource Centre for Consumer Participation in Health (NRCCPH). ⁹

One of the publications from the NRCCPH, ‘The report on consumer participation in resource allocation’, ¹⁰ is a comprehensive summary of methodologies for eliciting consumers’ views and involving the public in health‐care resource allocation. ¹⁰ In spite of the Commonwealth Government's commitment to encourage a ‘stronger, more active role for consumers at all levels of the health system’ ⁹ funding from the Commonwealth Government for the NRCCPH ended the 30th of June 2004.

There are published studies on community participation in decision‐making and priority setting in Australia. ⁷ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ However, to date there have been no published studies in Australia regarding public involvement in decision‐making priority setting for HCMs. This is despite the fact the public is considered to be a major stakeholder. ¹⁶ In an editorial regarding access to high cost, genetically targeted medications in Australia, Hall et al. considered that the economic and policy challenges imposed by these medications merit public discussion. ¹⁷ The importance of community consultation regarding access to HCMs can also be highlighted by the fact that one of the drivers of health‐care cost is considered by some to be the patients’ demand for HCMs. ¹⁸ , ¹⁹

The focus of previous research has been on the perceptions, attitudes and concerns of health‐care decision‐makers’ regarding access to HCMs. ⁵ The views of members of the general public are less well understood. Mooney and Blackwell expressed that citizens’ values are more likely than patients’ values to include interests in others in the society. ‘The community has a broader vision, more detached, more social and more considerate of the needs of others.’ ¹⁵

Therefore, it is appropriate that the perspective of the broader community is sought to inform future programmes regarding access to HCMs. Hence, the aim of this study was to gather information about views of members of the general public, regarding the allocation of resources to HCMs in the public hospital setting. This study is part of a broader research programme on issues in decision‐making about HCMs within the Australian health‐care system, aspects which have been reported elsewhere. ⁵ , ²⁰

Methodology

Selection of method

Different methods have been described to explore views of the public about allocating resources for health care. ¹⁰ These include: individual interviews which can provide detailed information especially about individuals’ experiences. Focus groups are useful to gain insight into people's views and answer questions about ‘why’ they think the way they do and ‘how’ they are affected. However, surveys are useful when exploring ‘what’ members of the general public think about a topic. ¹⁰ They also allow a wider cross‐section of the general public to be approached.

Sample selection

The sampling frame for this study was people living in the Sydney metropolitan area. The population in Sydney comprises 21% of Australia's population. ²¹ To be eligible, respondents needed to meet the following inclusion criteria: over 18 years of age, an Australian citizen or permanent resident, and able to complete the questionnaire in English.

Sample size

The target population was people in Sydney, and according to the 2001 census, this included 3 997 321 people. ²¹ For this study the level of confidence was set at 95%. The sample size calculation assumed a population diversity of 50/50 on variables (such as sex). A split of 50/50 is recommended as it produces the largest sample size requirements. ²² The margin of error tolerated for this study was set at ±7% (sampling error).

Questionnaire development

A literature review of previous studies exploring views of the public about allocating resources for health care identified potential items for the survey instrument. ¹³ , ¹⁴ , ²³ , ²⁴ , ²⁵ From this review a draft instrument was developed and modified to include criteria previously mentioned by health‐care decision‐makers. ⁵ The National Resource Centre for Consumer Participation in Health (NRCCPH) also provided advice. ²⁶ The final survey instrument contained 16 items divided into four sections.

Section 1 asked participants about their health status and use of public hospital services. The second section related to knowledge regarding public hospitals and paying for HCMs. People were given statements and asked to indicate if these were true or false.

The third section included six questions (Q11–16). Respondents were asked to allocate resources to HCMs in public hospitals given a limited budget. The first question in this section (Q11) asked respondents which factors were the most important in deciding who should be given a HCM. Respondents were asked to rank factors from 1 to 4, with one being the most important factor. This was then followed by two hypothetical scenarios with multiple options (see Appendix 1). Scenario 1: was the first hypothetical scenario. Respondents had a limited budget for medications in a public hospital. They were asked to provide a basis for choosing between two patients. In the second hypothetical scenario (Scenario 2) respondents had a limited pool of money and they had to spend it on two treatments: Medication A, to prevent heart attacks or Medication B to improve the quality of life of a person with cancer and lengthen that person's life. An open ended question was included asking participants their reasons for their response to the second scenario. In this same section (Q14–16) respondents were asked who should make those decisions, if they would like to be involved in the decision‐making and if they were willing to pay more taxes to subsidize access to HCMs in public hospitals.

The fourth and final section consisted of eight questions to collect demographic data which included: age, sex, marital status, language spoken at home, economic activity, highest level of schooling, postcode and annual household income. Only responses to sections 3 and 4 (Q11–13) are reported in this paper.

Pilot testing

The questionnaire was pilot tested with a convenient sample of 15 people. These were members of the community who met the eligibility criteria. Respondents were advised that the purpose of the pre‐test was to improve the questionnaire and were asked to critically evaluate it. Minor changes were made to the format as a result of this pilot testing.

Data collection and analysis

This was part of a survey about the community's knowledge and views regarding funding for HCMs in public hospitals and the criteria used in decision‐making regarding access to these medications. The survey was conducted between October 2004 and April 2005. Individuals were approached by the interviewer in venues such as shopping centres and on public transport in the Sydney metropolitan area. To make it a representative sample, people were approached at different times during the day and different days of the week including the weekend. ²⁷ Every fifth person approaching the interview setting (train station, shopping centre. etc.) was invited to participate. If they agreed the interviewer then supplied a subject information statement which provided background information and guided respondents through the questionnaire. Respondents were encouraged to express their views freely and told that it was their ‘views and perceptions’ that were being sought. Any issues could also be discussed openly with the interviewer. They were advised that data would be treated in confidence and all the information would be de‐identified.

Responses were collated and analysed using the Statistical Package for the Social Sciences (spss) for Windows Version 10 (SPSS Inc., Chicago, IL, USA). Descriptive statistics were used to summarize the data. Frequencies and proportions were calculated with their 95% confidence intervals. The sample demographics were compared with the Australian population as a whole using the one sample t‐test. To test the associations between responses and demographic characteristics (e.g. age, gender, income, etc.), responses were dichotomized and chi‐squared test was conducted. Significance was set at the 5% level.

Responses to the open ended question (scenario 2 –What are your reasons?) were imported into QSR NVivo® Version 2.0 (QSR International, Doncaster, Australia), qualitative research software. Responses were collated and analysed by identifying key themes.

This study was approved by the Human Research Ethics Committee of The University of Sydney.

Results

Section 4 – Demographics

Two hundred and ninety‐eight people were approached and 200 people completed the questionnaire. The reasons people gave for not accepting the invitation to complete the survey included: ‘not interested’, ‘don't have time’ or ‘couldn't read English’. Fifty‐six percent of the respondents were female and the median age was 36 years (range: 18–82). The sociodemographic details of the respondents are shown in Table 1 along with the 2001 census data for the Australian population ²⁸ and 2001 National Health Survey Data. ²⁹

Table 1.

 Sociodemographic characteristics

Socioeconomic characteristics	Sample population	Australian population
Gender (n = 200)
Female	56.0%	50.6%
Age (n = 199)†
Median	36 years	34 years
Standard deviation	14.0	22.5
18–24 years (n = 24)	12.0%	9.4%
25–44 years (n = 108)	54.0%* (CI: 47–61)	29.8%
45–64 years (n = 58)	29.0%	23.0%
65 years or more (n = 10)	5.0%* (CI: 2–8)	12.6%
Marital status (n = 198)†
Married	46.5%	50.7%
Language spoken at home (n = 200)
English	84.0%	80.0%
Level of education (n = 200)
University degree	61.0%* (CI: 54–68)	9.7%
Economic activity (n = 200)
Working (full‐time or part‐time)	88.0%	92.6%
Annual household income (n = 186)†
>A$100 000	54.0%	‡
PHI (n = 196)†
Yes	68.0%§* (CI: 61–74)	43.0%§

*One sample t‐test (P < 0.05).

^†Answers to some questions were not provided by all participants.

^‡Average household income in 2004 was $50 000. ⁴⁷

^§People covered by Private Health Insurance (PHI) refers to anyone with hospital cover, including those with hospital cover only and those with hospital and ancillary cover. ²⁹

The study sample differed from the Australian population in a number of characteristics. There was a higher percentage of females, a higher proportion of people between 25 and 44 years (P < 0.05), a higher proportion of people who were university educated (P < 0.05) and the sample was more likely to have private health insurance (PHI; P < 0.05). In addition, annual household income was higher than in the Australian population. It is important to note that Sydney has the highest average household income in Australia reported as A$63 000 in 2001. ³⁰

Section 3 – Factors and choices

Factors

The most important factors that respondents thought should be considered when deciding who should receive a HCM have been listed in Table 2. Respondents were asked to rank factors from 1 to 4, with one the most important factor. Most respondents (35%) considered that ‘treatment outcome’ of the recipients was the most important factor. This was followed by current health status (26%). Quality of life came third with 15% and the fourth place was for life expectancy and age.

Table 2.

 The most important factors considered when deciding who should receive a High Cost Medication (HCM)

What factors are most important in deciding who should get HCM?	Percentage (n = 196)	Rank
Treatment outcome	34.7 (68)	1
Current health status	25.5 (50)	2
Quality of life	14.8 (29)	3
Life expectancy	9.2 (18)	4
Age	9.2 (18)	4
Socioeconomic status	4.6 (9)	6
Family commitments	1.0 (2)	7
Lifestyle	0.5 (1)	8
Other	0.5 (1)	8

‘Other’ represents one respondent who wrote: ‘people who should have the most access to science, are those individuals who have dedicated most to elevate humanity for example like Ghandi’.

Scenario 1

In the first hypothetical scenario respondents had a limited budget for medications in a public hospital. They were asked to provide a basis for choosing between two patients with exactly the same characteristics. Eighty percent of respondents favoured a choice based on ability to benefit in terms of quality and length of life. Table 3 shows how respondents decided who should receive the treatment. A small proportion of respondents 2.5% (six) provided ‘other’ answers, these included:

Table 3.

 Who should receive treatment – Scenario 1

There is a limited budget for medications in a public hospital. Two patients need to be treated with a High Cost Medication for cancer but there is only enough money to treat one of them. How do you think the hospital should choose which one receives the treatment?	Percentage (n; n = 199)
One of them randomly	5.5 (11)
The one who benefits the most in terms of quality and length of life	79.5 (159)
The youngest	1.5 (3)
The one whose work contributes more to society	1.5 (3)
The one who has more family members to support	4.5 (9)
Don't know	5.0 (11)
Other	2.5 (4)

Every one should be entitled to medications and treatment no matter what the circumstances (Respondent #69). 
Couldn't choose – they both should get the treatment ( Respondent 116). 
Combination of all of the above except for one of them randomly (Respondent #152). 
Why do we have a limited budget? (Respondent #174). 
Other: A mix of factors weighted – there is no one answer (Respondent #179).

Scenario 2

In the second hypothetical scenario respondents had a limited pool of money and they had to spend it on two treatments: Medication A, to prevent heart attacks or Medication B to improve the quality of life of a person with cancer and lengthen that person's life. More than half of the respondents split the resources evenly between Medication A and B. Eleven percent allocated all the resources to Medication A and 5% to Medication B. Fifteen percent decided to allocate more resources to Medication A compared with B. Table 4 shows how people decided to spend the limited pool of money.

Table 4.

 Distribution of resources for Medications A and B – Scenario 2

There is a limited pool of money to be spent on medications. Medication A costs $40 per patient per month and could prevent heart attacks. Medication B costs $4000 per patient per month and could improve the quality of life of a patient with cancer and lengthen that person's life. What would you do?	Percentage (n; n = 194)
Spend all the money on Medication A	11.0 (22)
Spend all the money on Medication B	5.0 (9)
Spend some money on Medication A for 500 patients and some of the money on Medication B for 5 patients	66.0 (133)
Spend some money on Medication A for 800 patients and some of the money on Medication B for 2 patients	15.0 (30)
Other	3.0 (6)

The other answer (3%) was given by six respondents. Two respondents wanted more information or thought information was insufficient to make the decision.

Don't know – not enough information (Respondent #186).

Another respondent stated that:

Every one should be entitled to medication treatment no matter what the circumstances (Respondent #69).

The remaining three respondents thought money should be spent:

50% each (Respondent #47). 
Spend all of the money on patients as needed until it is gone then stop treatment with both drugs (Respondent #71). 
Would depend on demand (number of patients for each condition) (Respondent #115).

Among the 200 returned questionnaires, 117 (58.5%) had a written response to the open ended question to Scenario 2 – What are your reasons? To allow the reader to judge the veracity of the interpretation, quotations were used to illustrate the themes presented.

Utilitarian view‐point

The majority of the respondents perceived this was the most ‘equitable’ distribution of resources. ‘fairness’ and ‘solidarity’ were also expressed in the respondent's comments.

One needs to bear in mind ‘the common good’ ie balance benefits = ethical decisions (Respondent #11). 
All people have the right to life. Equal opportunities to live (Respondent #12). 
Society has a responsibility for all its members. After all, it is empathy and compassion that really separates the civilized from the uncivilized. However a cost/benefit analysis must also be made (Respondent #30). 
A skewed sense of responsibility and fairness re God like decision (Respondent #117).

Prevention over treatment

Those who decided to spend more money on Medication A expressed that with limited resources this would benefit more people. Some also expressed that they would rather spend money on Medication A to ‘save lives’ than B that only ‘prolongs life’.

The more patients that can benefit from the limited resource the better (Respondent #14). 
Ultimately you must save as many lives as possible but some exception should be made for the more expensive meds/improve quality of life (Respondent #29). 
If a medication costs $4000 they need to find an alternative way or a different medication. Is the drug really worth someone's hard earned money? I might change my mind 20 years from now (Respondent #32). 
Public health resources need to be allocated in a way that attempts to meet the health problems of a large group (Respondent #38). 
Can help more people possibly save lives. Whereas cancer drugs are merely prolonging life (Respondent #60). 
Cardiovascular disease is a bigger problem than cancer – more prevalence, effects more people. Scenario suggests that person/people with cancer are not going to be cured (Respondent #132).

Blame – personal responsibility

Respondents who decided to allocate more resources to Medication B believed that heart attacks could be prevented with lifestyle changes.

To my understanding people have no choice when they are struck down with cancer. Heart attacks can be mostly attributed to poor life choices and mostly at a later stage in life (Respondent #48). 
Life style, diet, exercise etc could alter/improve group A chances of not having heart attack whereas group B already have a life threatening condition (Respondent #93). 
Prevention is more than medication. Lifestyle factors are more important than drugs in this scenario (Respondent #171).

Access to all – no need to set priorities

Some respondents also stated that everybody should have access to these medications and choices should not be made.

Everyone should have some access to all medications (Respondent #82). 
All patients with different treatment needs have the right to receive the right treatment (Respondent #113). 
Couldn't choose‐they both should get the treatment (Respondent #116). 
… don't imply agreement with a system that would force such decisions. Politicians need to be accountable for their management of health policy and compromises that flow on from their decisions on resource allocation (Respondent #145). 
I believe it would be unethical to deny all public patients access to HCMs. Distribution should be to those who benefit the most in terms of efficacy of the medication (Respondent #162). 
It doesn't matter which answer is correct, the hospital should have enough money to supply both drugs (Respondent #176).

Involvement in decision‐making

Half of the respondents (50%) said they would not like to be involved in decisions regarding access to HCMs in public hospitals (Fig. 1). Gender, income, economic activity, level of education and health status were not predictors of respondents’ involvement in decision‐making (see Table 5).

Figure 1.

 Willingness to be involved in decisions regarding access to High Cost Medications in public hospitals.

Table 5.

 Predictors of respondents’ involvement in decision‐making

Characteristic	Willingness to be involve
	Yes, n (%)	No, n (%)	χ 2*	P‐value
Gender
Female	46 (61.3)	51 (50.5)	2.044	0.170
Male	29 (38.7)	50 (49.5)
Income
<30 000–49 999	18 (26.1)	32 (33.0)	0.913	0.393
50 000 >	51 (73.9)	65 (67.0)
Economic activity
Employed	71 (94.7)	86 (85.1)	4.049	0.051
Unemployed	4 (5.3)	15 (14.9)
Level of education
High school	7 (9.3)	20 (19.8)	3.632	0.061
>Certificate	68 (90.7)	81 (80.2)
Health status
Good health	67 (90.5)	88 (87.1)	0.491	0.632
Poor health	7 (9.5)	13 (12.9)

*d.f. = 1.

Discussion

Factors

Respondents considered treatment outcomes, current health status, quality of life, age and life expectancy as important factors in deciding who should have access to HCMs. Similar results have been found elsewhere. These results refer to individual circumstances where ‘benefit to patient’ and quality of life are considered factors to allocate resources. ²⁴ In the United Kingdom, a survey regarding health‐care rationing showed that most respondents agreed or strongly agreed that patients’ quality of life should be considered. ³¹

Even though factors were identified there was no attempt to make a quantitative trade‐off between them. Conjoint analysis could be a useful way to evaluate trade‐offs between factors as well as discrete choice experiments. However, this study aimed to explore and not to quantify.

Scenario 1

The majority of respondents (80%) considered that the person who benefits the most in terms of quality and length of life should have access to a HCM in scenario 1. In 1998, the Eurobarometer, a survey which is meant to monitor the public opinion in the European Union, had a section on rationing and priority setting. One of the questions asked participants to choose between two patients with chronic heart disease and exactly the same characteristics (age, employment, family support). Consistent with the results of our survey the European response was ‘whoever can benefit most in terms of quality and length of life’. ²⁴

One of the limitations of scenario 1 was that length of life and quality of life were not given as separate factors. Therefore, it was not possible to determine the importance respondents may have given to prolonged life vs. improving quality of life or the trade‐off between both of them. However, there was a need in constructing the questionnaire to balance between simplicity and completeness. Respondents already had five options to choose from and giving them another extra two options would have made it even harder to deal with. However, the results presented in Table 2 show that participants ranked quality of life more highly than life expectancy.

Scenario 2

In scenario 2, most respondents’ allocated resources according to what they considered would benefit the greatest number of people. Respondents stated that they tried to divide the limited amount of money equally across disease states. The concept of fairness implied by respondents meant that both groups of patients would benefit from a treatment. For those who gave priority to Medication A over B, prevention was regarded as important. Medication A was favoured on the grounds that Medication B was not going to provide a cure. Those who decided to choose Medication B over A stated that medication is not the only way people can prevent heart attacks. Respondents mentioned that a healthy lifestyle may also prevent heart attacks. These results were similar to those of Alexander's study. ¹³ In that study participants chose services that would benefit the greatest number of people. Respondents also considered that money should be invested in services which diagnose illness and provide early treatment. Services which aimed to prolong individuals’ life had a lower ranking. ¹³

For Williams, those people who benefit the most from the resources available should be the ones that receive priority. ³² The values of the community as a whole must override the values of a particular interest group within it. ³²

Involvement in decision‐making

Half of the respondents (50%) did not want to be engaged in decisions regarding access to HCMs in public hospitals. This result is consistent with previous studies. ¹⁴ , ²⁵ In a study conducted by Litva et al., the role of the public was seen as limited regarding decisions concerning choices between patients. Participants were asked if they wanted to be involved in the decisions about which patient gets a medication within a limited budget. The majority of the participants were unwilling to have any role in this type of decision. ²⁵ In Wiseman's study, patients’ were also less willing to be involved in decisions regarding ‘specific medical procedures’. ¹⁴ According to Litva et al., the willingness of the public to participate is based on the level on which priorities are being set. ²⁵ It might be concluded that the public is less willing to participate in the allocation of resources at the institutional level. However, they could be involved as expressed by Litva et al., in assisting to set criteria for deciding between potential beneficiaries of treatments. ²⁵

A frequent comment by respondents was that the questions about factors and scenarios were difficult. The scenarios were criticized as respondents believed they should never occur. This suggests that some respondents were unaware of the extent to which rationing decisions are made. Some authors have previously noted that health‐care rationing is not explicit in Australia. ³³ , ³⁴ , ³⁵

King and Maynard, in the UK, contended that a debate is necessary for greater awareness among all citizens of issues involved in rationing and priority setting. ²³ The Australian public has not been exposed to issues of rationing and priority setting and therefore might have less familiarity with the issues.

Demographics

Surveys of members of the Australian community about priority setting in health care have been conducted previously. ¹² , ¹⁴ The sociodemographic characteristics of the respondents in those studies appear similar to the respondents in this study. Both Wiesman and Nord also reported that, compared with census data, respondents were more likely to have private health‐care insurance. ¹² , ¹⁴ The differences between the respondents in our study and the census descriptions are noted. Furthermore, it is acknowledged that people who were illiterate or had difficulty reading were visually impaired and those who speak languages other than English were excluded from this study. ³⁶ Even though the sample was not representative of the Australian population at‐large, the results do provide an insight into the views of the community about access to HCMs in public hospitals.

Baume states that the general public has insufficient knowledge to play a useful role in making decisions about hospital or medical resource allocation. ³⁷ However, it is the community that bears the cost of health‐care allocation decisions as taxes pay for these resources. ³⁷ Several authors have suggested that the values of the community may be different from those of politicians, health‐care providers and health‐care professionals. ¹⁵ , ³⁷ , ³⁸ , ³⁹ From this perspective it is necessary to inform the public about how the health‐care system is funded in Australia, including the mechanisms for access to medicines, and to seek the considered views of at least a representative sample of community members. Smith and McGettigan considered that an informed consumer sector has the capacity to influence government and add transparency to national committees. ⁴⁰

Those who support the involvement of the general public in decision‐making priority setting consider that it increases the local accountability and decision‐makers should be accountable to the actual and potential consumers. Decisions should be made more explicit and open. ²⁵ For Lofgren it is no longer acceptable for decisions regarding access to pharmaceuticals in Australia to be taken ‘behind close doors’. ⁴¹ Lindblad et al., in Sweden, considered that involving patient perceptions in the priority setting process for HCMs might facilitate the implementation in practice. ⁴²

Limitations

This survey did not choose participants randomly, which could have resulted in selection biased. As previously discussed the sample differed in a number of characteristics from the Australian population.

Comparison of data from surveys can be challenging. Differences in study designs, samples, questionnaire wording and analyses sometimes makes it difficult to compare results across studies regarding allocation of resources in health care. ²³ , ²⁴ , ⁴³ It is difficult to interpret results from questions which are not identical. There are problems and ambiguities between surveys. ²⁴ Caution should therefore be taken when comparing results. The majority of the surveys have looked at the general public and allocation of resources in health care and this survey looked at the allocation of resources to HCMs in public hospitals. In a review of studies involving citizens in rationing Mossialos and King expressed that ‘there is considerable ambiguity’ regarding ‘very expensive’ high tech treatments. Also, data derived from surveys sometimes differ from focus group discussions. ²⁴

One of the limitations of surveys as previously stated is that the wording of questions and the way they are presented influences responses. ³⁶ , ⁴³ , ⁴⁴ Finally, surveys such as the one undertaken for this study could be criticized for not allowing respondents to deliberate due to the limited amount of time and information. ⁴³ Bowling et al. wrote that it is important to provide the public with the necessary information to make informed judgements. ⁴⁵ Mooney and Blackwell suggest that respondents need time to reflect and deliberate. ¹⁵ Nevertheless, despite these limitations, surveys allow a wide cross‐section of the public to be approached. ²⁴ They also form one body of evidence that can be supplemented by other methodologies such as focus groups and interviews. ⁴⁶

This study was conducted to gather views of members of the general public about access to HCMs in public hospitals. This information addresses a gap in the knowledge needed to make informed decisions about the allocation of resources to HCMs in public hospitals. Health‐care decision‐makers make choices between patients and/or treatments in resource constrained environments. However, there is a lack of adequate guidance for those decisions based on community views. Mooney and Blackwell wrote that ‘health‐care spending decisions made within budgetary constraints are not generally made in close consultation with the community’. ¹⁵

From these findings it is clear that the majority of the public do not want direct involvement in decision‐making, although some do. This could mean that decision‐making processes could directly include interested members of the community. Alternatively these findings offer support for the development of a process for involving community members in discussion on policy on the provision of treatment and services within health‐care institutions and seeking the views of members of the public on the provision of these HCMs and other selected expensive services within public hospitals, rather than involvement in every decision which has to be made.

Appendix 1. Section 3 (factors, scenarios 1 and 2)

Factors provided to participants to be ranked:

• • Age
• • Socioeconomic status
• • Current health status
• • Life expectancy
• • Quality of life
• • Family commitments
• • Lifestyle
• • Treatment outcomes
• • Other (specify)

Scenario 1

There is a limited budget for medications in a public hospital. Two patients need to be treated with a HCM for cancer but there is only enough money to treat one of them. How do you think the hospital should choose which one receives the treatment?

By Choosing:

[Please tick one box only]:

•  one of them randomly

•  the one who would benefit most in terms of quality and length of life

•  the youngest

•  the one whose work contributes more to society

•  the one who has more family members to support

•  don't know

Scenario 2

There is a limited pool of money to be spent on medications. Medication A costs $40 per patient per month and could prevent heart attacks. Medication B costs $4000 per patient per month and could improve the quality of life of a patient with cancer and lengthen that person's life.

What would you do?

[Please tick one box only]

•  Spend all the money on Medication A for 1000 patients.

•  Spend all the money on Medication B for 10 patients.

•  Spend some of the money on Medication A for 500 patients and some of the money on Medication B for 5 patients.

•  Spend some of the money on Medication A for 800 patients and some of the money on Medication B for 2 patients.