The state of health services partnering with consumers: evidence from an online survey of Australian health services

We took a whole of population approach and used the Australian Hospitals and Aged Care Database to identify services (see http://​jpmmedia.​com.​au/​direct-marketing/​hospitals-aged-care-database/​about-hospitals-aged-care-database/​). To ensure questionnaires were directly sent to the most suitable participant, we telephoned all health organisations to identify an appropriate personal, rather than general organisation, email address. Criterion for inclusion was that the participant was either a) the CEO or equivalent of a participating health service, b) CEO representative (nominated by the CEO), or c) the person in charge of accreditation (and the NSQHS Standard Two specifically) for the particular health service. One participant/return only, from each emailed address was requested to avoid multiple responses from the same service. While compiling the sample, we found some CEOs have responsibility for multiple health service organisations. We noted these in our database and decided to send each of these CEOs only one email. This was to avoid confusion about distinguishing between health services when surveys were returned. After deducting duplicates, health services that had closed or merged, and those covered by a ‘parent’ health service organisation responsible for accreditation, we identified a total of 447 individual participant email addresses, covering 1200 health services in total.

The survey (see Additional file 1) was designed to collect data to address the two research questions. Demographic data, including about the state or territory location, respondent job title, and service type, were collected. For type we devised a classification aligned with Australian health services that provide hospital or day surgery services: i.e. publicly-funded, publicly-funded community health service, metropolitan, regional, sub-regional, small rural, multi-purpose, community-based, private hospital or other (asking respondents to specify). We devised this typology based on testing of survey drafts with four health service managers. Australian states have different categorisations for health service types so we devised a typology intended to be useable across Australia.

Questions to address Research Question 1 followed the activities for consumer partnering identified within domains of NSQHS Standard Two; for example, within the domain ‘service management and evaluation’, questions included: “Have you used consumers in the analysis of safety and quality performance information and data, to develop plans for your health service?” and “Have you partnered with consumers in the analysis of patient feedback data?”. Responses could be: yes, no or don’t know, followed by a request for explanation (e.g. briefly list how you have done this). To address Research Question 2, we included a set of minority groups likely to have particular needs or perspectives based on their experiences. We based these on categories from a NHS England document [16]. We asked respondents if their services had “specifically sought to include” each of the specified minority groups. The survey asked how health services recruited consumers from these groups; or if not, then why. Finally, a box was provided where respondents could record anything else on the topic that they wanted to say.

Survey questions and format were loaded into Qualtrics. A link to the survey was sent by email to the identified participants. The email included an introduction letter, plain language information sheet and ethics statements. Once the plain language statement was read, the survey took about 15 min to complete. Consent was obtained via response to the survey. All surveys were returned anonymously via Qualtrics. The survey was first emailed in September 2016 (69 surveys completed), with follow-up reminders in October 2016 (43); and two in November 2016 (33 and seven). We undertook to send all participants a findings summary at study end.

Quantitative survey data were input into SPSS (IBM Version 23) and exploratory analysis conducted to examine respondent characteristics and frequencies of responses. Simple cross tabulations and Chi-square tests were calculated to determine associations between health service type (private versus public) and geographic location (metropolitan versus non-metropolitan) and service engagement in consumer partnership activities. Qualitative responses were input into NVivo (Version 11). These tended to be brief and note-like rather than discursive, due to the nature of the survey response format requested (i.e. ‘please list…’). Inductive analysis was conducted to identify thematic categories. Data were then allocated to thematic categories and frequencies of quotes on themes counted [28]. KC conducted initial qualitative analyses. The other authors also read the data. JF checked and verified the analysis.

Table 2 shows the number of all health services that stated they were engaged in the consumer partnering activities covered in the questionnaire (i.e. that responded yes – their service has undertaken this activity). Responses are broken down by type (private and public) and geographic location (metropolitan and non-metropolitan). Due to the nature of the sample, none of the health services are specifically primary care services, although some of the services may provide primary care as part of their service portfolio. There are no specific Aboriginal and Torres Strait Islander hospital or day surgery services, but some hospitals responded that we can infer would have high use by the Aboriginal and Torres Strait Islander community.

Exploratory descriptive analysis showed that, overall, the majority of respondents reported their health services were engaged in partnering activities. The proportions of services engaged in these activities ranged from 95.6% of respondents partnering with consumers ‘to develop or provide feedback on patient information’, to 61.3% partnering with consumers ‘to develop information about health service safety and quality performance’. Chi-square tests of association showed a significant association between type of service - public - and engagement in: strategic and operational services planning, governance structures, service design and analysis of safety and quality performance data (in bold see Table 2). Chi-square tests were calculated to determine if there was an association between metropolitan or non-metropolitan location, and partnering activities. Analysis showed only one significant association between location of service – non-metropolitan - and engagement of consumers in analysis of patient feedback data (in bold see Table 2).

The survey asked about inclusion of people from minority groups in consumer partnering. Descriptive analysis indicated that specific steps to include the minority groups listed occurred for less than half of responding services (see Table 3). People from different cultural and ethnic backgrounds was the most frequently reported group that services sought to include (45.5%). Socially disadvantaged people was the least reported group (23.6%). Responses are broken-down by type (private and public) and geographic location (metropolitan and non-metropolitan). Chi-square tests showed a significant association between service type - public - and services seeking to include all of the minority groups described, except people of different genders (in bold Table 3). There was a significant association between location – metropolitan – and seeking to include people from different cultural and ethnic backgrounds and people from the LGBTI community (in bold Table 3).

As noted previously, due to the nature of the questionnaire format (where participants were asked to ‘list’, for example, activities or reasons for doing/not doing, activities), qualitative responses tended to short, note format, rather than discursive format. This type of response lent itself to compiling frequencies of responses by theme, rather than writing summaries of each theme. Table 4 thus provides frequencies of partnering mechanisms reported (see Table 4); and Table 5 descriptions of partnering activities, along with verbatim exemplar illustrations of items listed.

The most common mechanisms listed were community advisory committees and consumer representation on other health service committees. The most common activities of community advisory committees were reviewing and providing feedback on patient data, safety and quality data and patient information. A small number of comments indicated that some health services included their advisory committee in service decision-making. The committee that consumers were most likely to attend as representatives was ‘quality and risk’/ ‘quality and safety’.

There were 160 qualitative responses to questions about including people from minority groups. Table 6 summarises the ways respondents “sought to include” participants, with illustrative comments. Some respondents noted more than one strategy. The most frequently reported active recruitment method was through other groups or services in the community. Other ways included approaching service users directly and through specific health service teams.

Table 7 summarises the reasons given by respondents about why they did not seek inclusion of minority groups. A total of 276 responses were received. Some gave more than one reason. Some respondents considered it irrelevant or inappropriate for their organisation, while others suggested there are low numbers of people from minority groups in the local area, that they have limited resources to seek out participants, or there was a lack of interest from relevant groups.

The majority of services are engaged in partnership with consumers (95%) and used community advisory committees as a key mechanism. One of the requirements of the NSQHS [3] in relation to consumer partnership is ‘establishing governance structures’. Forming a specific community advisory committee appears to be a relatively non-disruptive way for services to meet this requirement, by ‘bolting-on’ a new committee rather than taking steps to integrate consumers into existing committee structures and activities. This finding aligns with previous findings of the Australian Commission on Safety and Quality in Health Care [14] evaluation where involving consumers in committees was considered, by health services, to be one of the easiest consumer partnership activities.

Through involvement in a specific consumer advisory committee or being placed as a consumer representative on other committees, our findings show that consumers are drawn into various activities. However, these could be interpreted as being in relatively superficial areas of activity - such as providing feedback on patient information. There are different ways to view this situation. Established frameworks for analysing citizen involvement might understand this activity as ‘consulting’ [29] or ‘participation as a means’ [30] and thus relatively tokenistic and instrumental; with deeper structural and more system-wide involvement necessary to producing the empowered citizens and communities that bring population health impacts [31].

However, Tambuyzer and colleagues [32] highlight that, in consumer partnering, ‘one size does not fit all’ [p145]. Participation does not always have to aim for complex activities and there should be awareness that consumers may not desire to be involved in service strategy formulation or governance. It may boost consumers’ confidence to work together as a group, rather than be spread as individuals across a range of activities. Tritter and McCallum [33] highlight that consumer partnering in health activities is a complex business – that different individuals will want to contribute in different ways – and that health organisations are best served if citizens provide a range of inputs from commentary to strategic and directional involvement.

This makes having a specific consumer committee a beneficial activity, but perhaps within a portfolio of optional activities for partnering. If a consumer committee is the main focus of activity, then this runs the risk of appearing to be an easily-established after-thought, bolted-onto rather than integrated into a health service organisation. O’Shea et al. [34] suggest that, if consumer partnering is about co-production of services rather than ‘box-ticking’ to say a consumer has been consulted, then health services should include consumers in a range of activities where their perspective is of equal value to health employees [34].

Just under half of responding services had recruited consumer participants from minority groups. Inclusion of diverse participants is often highlighted as challenging in consumer partnering [35]. Nonetheless, we found several health services actively applying strategies to include minorities, including approaching other groups and services, their own service users and even targeted advertising. This shows that inclusion of minority groups is an issue acknowledged as significant by some services that have implemented strategies to specifically attract minority groups. These services may be early adopters, providing useful experience from which others could learn. By contrast, some respondents noted that participants from minority groups were not actively sought out because they already attained diversity due to the demographics of their location.

Some service respondents said they did not recruit from minority groups because it was irrelevant. Perhaps this is justifiable for some specialised services or for services located in areas with homogeneous populations. However, given research showing that minority groups tend to be under-represented in consumer partnering [36], it seems important for services to have robust evidence justifying non-inclusion of minority group participants. Other respondents highlighted difficulties in attracting minority group participants; for example, people from the LGBTI community in rural areas. This highlights a challenge in identifying participants with particular experiences without causing stigmatisation – an issue previously highlighted by other studies (e.g. [37]). Overall, it appeared that compared to other minority groups, less efforts were made to include socio-economically disadvantaged people. This could be because it is difficult to ‘label’ and specifically identify socially disadvantaged people – for example, in advertising – because they are perceived to have more ambiguous distinguishing characteristics, compared with disability or ethnicity. That is, perhaps it is more culturally embedded in Australia, to identify (i.e. label as ‘other’) some minority groups, compared to others. There was less inclusion of minority groups in non-metropolitan areas, and this could be due to reticence to look for, identify and include minority group participants due to the close-knit nature of rural places [38].

One issue may be a lack of tools to inform Australian health services about strategies for inclusion of minority groups, including in different geographical settings where there may be fewer numbers of people belonging to minority groups. Technology is increasingly mooted as potentially useful for including diverse views – either to actively include diverse participants in conversations (e.g. using social media discussions); or establishing specialist discussion groups that can be followed on social media to engage with important topics [39]. The NHS England [16] guidance on engaging with minority groups, notes that giving multiple ways to partner is useful for engaging minority groups, with technology adoption as one useful option. However, considerable care has to be taken with considering technology as a solution because digital exclusion due to poverty, variable connectivity, illiteracy, communication challenges and limited cultural sensitivity still present considerable barriers [39]. Using technology did not arise at all as a strategy deployed by health services in our survey, which may suggest a gap for future study.

There was a significant association between public, as opposed to private, health services and consumer partnering activities; and a strong association between public health services and engagement with minority groups. This could be attributable to public services having to be more accountable for meeting public needs, as they are paid for by taxpayers. There is also evidence, particularly for small, non-metropolitan public health services, that they take considerable steps to build community inclusion and cohesion – so they are perhaps more culturally aligned with seeking inclusion of minority groups [40]. It could also be the case that public health services are more location-specific, embedded in the geographical communities, whose populations form strong connections to them. Private services tend to have wider reach, based more on volume of demand and patients’ capacity to pay than proximity to community, potentially making it more difficult to attract volunteer consumer partners.

Non-metropolitan services were more engaged in consumer partnering than metropolitan services, but were less likely to include minority groups. Greater engagement in consumer partnering activity may reflect a closer alignment with principles of primary health care provision among rural health services. Primary care strongly emphasises the role of community participation and empowerment [30] and some studies of Australian rural health services have highlighted a specific mission regarding community capacity-building [41]. Less inclusion of people from minority groups potentially reflects the demographics of Australian rural areas which are traditionally less culturally diverse than metropolitan Australia [42]. The challenges of highlighting ‘other-ness’ in rural places, for the potential stigma it brings, has already been noted above. It is perhaps more risky to try to identify diverse people in rural areas, more risky for them to be engaged, and due to smaller overall numbers, there are likely to be fewer participants from each of the minority groups.

The study has limitations. There was uneven response across States and Territories, with highest numbers of responses from the States of Victoria and New South Wales. In relation to numbers distributed, there was poorest response from South Australia with only 15% responding compared to a range of 21–33% for most other States/Territories. The uneven response means we know little about how the culture of State health service systems might influence consumer partnering.

The survey was limited to services providing hospital and day surgery services because these services are covered by the national standards. If primary health services and Aboriginal and Torres Strait Islander services were also specifically studied, there is potential that a range of different strategies and activities would be identified. A range of strategies and deep involvement of consumers were found by Freeman et al. [30] in a study of six Australian primary healthcare services, for example.

We restricted response options to ‘yes’, ‘no’ or ‘don’t know’. It might have been useful to provide a finer-grained scale so that we could establish the extent to which activities were being conducted, although some information was provided through the listing of activities by services – where they took the time to do this. Providing a scale or more nuance around explaining activities could be useful in future research of this topic.

Findings might have limited transferability beyond the responding group. It is known that people are more likely to respond to surveys if the topic interests them (i.e. selection bias) - either because they are particularly affected by, or interested in, the items asked about. As people who respond almost certainly have different characteristics than those who do not, the results may be biased [43]. If the respondents are those that are most active and interested, then the findings could inflate the extent to which consumer partnering occurs. Innovators in consumer partnering may have self-selected to respond. Participants self-reported and the survey only accounts for CEOs’ (or their representatives’) perspectives. It is also possible that findings may be subject to response bias – given that the survey concerned compliance with standards, it is possible that respondents’ subjective reports were more positive than an objective measurement would verify. Further research also considering perspectives of consumers and staff, at the same services as the CEOs, would be valuable in generating a more accurate picture.

Finally, a revised set of Australian NSQHS standards has just been published [44]. Regarding consumer partnering, these highlight consumer committees and consumer involvement in activities around health literacy as core partnering activities. They continue to highlight the importance of including minority groups. While in the aforementioned respects, the new standards appear sharper than previously, there remains a lack of clarity about the nature of partnering that health services should pursue beyond apparently easier to bolt-on, activities.

Given the limitations outlined, as an exploratory study - systematically examining for the first time, the extent of consumer partnering in Australian health organisations – we suggest the findings make a significant and novel contribution to the field.