Patient and public involvement in health literacy interventions: a mapping review

Plain English summary If people can read, understand and act on health information to better their health and reduce illness, they are thought to have “adequate” health literacy. Poor health literacy can mean people are less able to access health care and manage their health. Health literacy tends to worsen as adults get older, and is especially poor in adults age 65 and over. Ideally, health literacy interventions target people before age 65, to establish good skills and habits before people have many health problems associated with ageing. It is also good if researchers consult ordinary people, including patients and the public (PPI) when planning a programme to try to improve health literacy. This may help ensure individual needs are catered for. We therefore looked for studies that described any role of patient or public representatives in the research planning stages. We explored how the representatives contributed to each project. We found only 20 studies that included people other than the research team. Lack of reporting and consultation with patient and public representatives may contribute to less success when public health programmes are undertaken. Abstract Background Health literacy is the ability to understand, access and use health care and is a critical mediating factor that affects the health of older adults. Patient and public involvement in health and social care research, policy and design of care delivery is one mechanism that can promote production of better health literacy. This mapping review looks for and describes practices, concepts and methods that have been reported involving patients and public in the development and design of health literacy interventions for older people. Methods Studies for the present review were selected from an inventory of health behaviour studies published between 2003 and 2013. The inventory was created by systematic searches on bibliographic databases (Medline, CINAHL, Scopus, Google) for health literacy interventions involving older people (50+ years) and resulted in screening of 5561 articles, of which 1097 met study inclusion criteria. For the research described in this article 96 of the 1097 studies specifically focused on health literacy and were independently screened by two reviewers to assess involvement of stakeholders other than investigators and participants. Results Twenty studies included patient and/or public involvement in at least one research domain: design, management or evaluation. Involvement included volunteers, older people, patients, and/or community representatives. Conclusions Patient and public involvement were rarely reported in studies on health literacy interventions for older people. Future intervention development needs high quality PPI, which is well reported to develop the evidence base and inform practice.

might not understand the health information they are given, especially if not culturally specific, or they may wish to know how best to live with a chronic condition rather than what is the "best" treatment. In so doing, it should be possible to move away from health and social care research that is done to people to that being done with people [23]. However, as yet, few clinical trials include patient-reported outcomes [24]. Health literacy broadly covers knowledge, behaviour and health outcomes that differ at the individual or group level. Using PPI in health literacy interventions should, therefore, be key to improving health outcomes.
This mapping review attempts to find and describe PPI in a sample of health literacy interventions. The purpose of a mapping review is to categorise studies and types of investigation and explore linked concepts used in a body of related research [25]. First, we implemented a customised search strategy to try to find reports on health literacy interventions that might include PPI. Then we read articles carefully, selecting for further descriptive analysis, articles that reported PPI activity. We describe PPI features that were found in the eligible articles. The results are described qualitatively but we do not treat these observations as definitive.

Data sources and selection criteria
Previous work by the authors [26] created an inventory of health behaviour studies published in 2003-2013, with the aim of developing policy and practical guidelines for health literacy promotion in Europe. The inventory was originally created by searching bibliographic and grey literature databases and sources (Medline, CINAHL, Scopus, Google) using a broad range of search terms related to health literacy and self-efficacy skills, for all interventions which included any older adults (although explicit targeting of this age group was not essential; "older adult" = age 50+). Age 50 has been suggested as a reasonable threshold for European public health research on older adults [27]. To create the inventory, 5561 articles were screened independently by two researchers; inclusion was confirmed by a senior academic with experience in health literacy research. 1097 articles were included in the inventory from the initial eligibility criteria, which were the inclusion of older adults in health literacy or compatible interventions [26]. For the research described in this article, we screened the title and abstract of these 1097 articles contain the exact phrase "health literacy" (n = 96). The abstracts of these 96 studies were duplicate screened for descriptions of studies that seemed to possibly describe involvement of stakeholders other than patients and investigators. The full texts of any abstracts that could not be excluded were read to search for the involvement of non-investigator patients, carers, community or charitable bodies in any aspects of research design including design of delivery, monitoring of evaluation (NIHR, 2014). Studies were excluded if it was not possible to discern that any of these groups were involved in any aspect of the research process. Among the remaining studies that described PPI involvement, we next posed two questions: Who is involved in the research? When and how are they involved?

Analysis
We categorised patient and public involvement (PPI) within each study. Data were extracted for the number of people involved in each stage of the research process, who was involved and how were they involved. Following published guidance on possible PPI roles [28], we devised a taxonomy for PPI opportunities to input to the research process, designating the contribution domains (see Table 3) as follows: Identification and prioritisation of research objectives; (other aspects of ) Design; Grant development; Project management or undertaking; Analysis and interpretation; Dissemination; Monitoring or evaluation. Two experienced qualitative researchers (SHW & CS) independently extracted data with differences resolved by a balanced discussion. There was an option to refer to a third investigator for very difficult decisions.

Results
The study selection process is shown in Fig. 1. Of the 96 studies that mentioned health literacy, 43 were removed due to no suggestion of PPI in the title or abstract. Fiftythree studies were read in full and 20 studies fulfilled the inclusion criteria (evidence of PPI reported). Details of studies are shown in Table 1, with description of PPI.
Much of patient and public involvement in the 20 studies on health literacy was concentrated into two main domains, design and/or evaluation of the research ( Table 3). None of the studies involved patients or the public to identify or prioritise areas or concerns for research, although Aspinall et al. [29] conducted a needs assessment to identify topics for workshops to improve the health literacy of older people. In design, some studies used PPI to develop educational material in various forms and for different uses [29,31,33,37,43]. Three studies used PPI to inform development of websites [30,39,45] and another to educate marginalised groups on the use of health websites [38]. PPI was used to address the unmet health education needs of ethnic minorities [35,[40][41][42], by enlisting the help of community groups to design interventions that addressed cultural and language barriers. One study enlisted PPI in recruitment [40] and one developed a questionnaire to assess unmet needs of patients [48].
In the undertaking and/or management of research, PPI was active in assisting with educational classes [37,46], counselling [47], sharing experiences of illness [44], improving cancer screening [33] and disseminating information into communities [38]. Other studies found that information learned through community events was also disseminated to other community and family members [35,41].
Patients with relevant chronic health conditions were common representatives in research that included PPI. Most studies were conducted in the USA, where patients were involved with design [31,37] or evaluation of intervention tools [32,34,36]. In Europe patients evaluated self-care cards [43], assisted in research management by sharing their experiences in health education programmes [44] and were involved in design, management and evaluation of a website for self-help chronic lower back pain [45]. Patients, along with all the other groups assisted with design and evaluation of self-help educational material for Punjabi and Chinese asthma sufferers living in Canada [42]. American older people contributed to design [29,33] and evaluation [30]   Note: Many studies reported differing numbers of representatives at different stages of the research, thus were added to several categories. Please see Table 1 for full details of interventions. Another study included older people and Alzheimer's Association members in design and evaluation of an educational tool about dementia for Spanishspeaking older Latinos [41]. Older people in Australia assisted with design of a tool to assess concern about asthma [48]. In the UK, older people, together with participants from most other groups (community/volunteers/family) contributed to design, management and evaluation of a fitness intervention to improve health behaviour [46]. Three more American studies included minority (ethnic and lower socioeconomic) group members, two in design of cancer interventions [35,40], and one for design, management and evaluation of health promotion that also included an advisory group [38]. In one European study peers spoke of experiences of their health conditions to people with similar conditions [44]. One study engaged family members in the development of ethnically and culturally suited educational material for asthma self-management [42]. However, family members were cited by older people as important sources of support for lifestyle changes [46] and for understanding medical information [29]. Interestingly it was not until the evaluation stage that one study found out how important families had been in helping patients use websites to improve health literacy [36] (Table 5).

Discussion
The aim of this mapping review was to consider when and how PPI activity has been reported in the research process within health literacy studies. A mapping review is an appropriate method where evidence is known to be difficult to find and describe, also when the conceptual definitions and boundaries are relatively new or unclear [25,49]. A mapping review enabled us to consider some emerging evidence in the nascent and wide ranging field of health literacy intervention research. Twenty diverse studies were found that reported PPI in the research process. Most studies were conducted in the USA where many addressed health literacy of minority groups. There appears to be a large gap in knowledge about health literacy for minority groups in Europe and other developed countries. PPI was used minimally in most studies. Interventions in Europe (including UK) were more likely to use PPI more comprehensively than those in the USA. Most of the studies in the USA involved PPI at only one stage of research, usually design or evaluation. For example, Aspinall conducted Note: Several studies reported participants at different stages of the research, thus were added to more than one category. Please see Table 1 for full details focus groups to identify needs assessment to build an educational program to improve health literacy [29] and in another study patients were recruited to evaluate a website designed to help with diabetes care [36]. Mayberry's intervention did little for health literacy and the researchers realised the importance of support given by family members only on evaluation. This example highlights the need for PPI to be inclusive of stakeholders that are involved early in the research process. In contrast, Schulz included patients in design, management and evaluation of a website offering self-help to those with chronic back pain [45]. Interestingly, Long only involved patients with diabetes for evaluation of the intervention. However, the intervention was successful in improving knowledge, behavioural and health outcomes, because telecare call handlers provided time and space for patients to develop rapport and ask questions about their condition and related subjects [47]. Patients found it more helpful hearing about living with a chronic condition than about the diagnosis from peers [44] demonstrating the importance of understanding what people want and need to know. Interpersonal relationships and communication issues throughout the health system are important factors to enable patients to understand their health conditions [17]. Many patients are confused by the language used in medicine and further hindered by ill-health that drains them of energy, although others may become experts on their condition over time [17]. Health literacy affects patients' and carers' ability to actively take part in their healthcare [3]. However, few people have any concept of health literacy and for many the term reflects academic language acquired in higher education [17]. Understanding the needs of people to improve their health should drive health literacy interventions. Reflections reported in some of the studies showed that interventions addressed unmet needs [47,48], improved patients' self-confidence and ability to communicate with medical professionals [45], and that understanding cultural values had educated a community about dementia [41]. These examples show the wideranging nature of the concept of health literacy and suggest that high quality PPI is required to understand and address people's needs, which in turn, could play an important role in improving health literacy and other interventions [17,24,50].
The studies included in our review showed no evidence of PPI living up to the 'gold standard' (e.g., GRIPP2) covering the core six principles proposed by Wilson et al. [20] and Staniszewksa et al. [51]. For example, no study reported formally on having independent patient or public members to sit alongside the researchers and feed into the research process. Most studies that included PPI did so at design stage, but none of the studies apparently elicited input at the inception of the research idea. A key finding was that overall PPI was often poorly reported, which aligns with results found by Note: Several studies reported differing numbers of PPI representatives at different stages of the research. Please see Table 1 for full details others [50]. Transparency is a much needed requirement and recommended as a national standard for PPI [52]. Lack of transparency and poor reporting reduces propensity to replicate the research, evaluate the actual usefulness of PPI and to explain the factors influencing the outcomes. In some studies patient representation was weak as it was only used for evaluation [53]. Practical problems linked to PPI input to research development were sometimes discussed. For instance Coughlan et al. [43] used members of the Pharmaceutical Care Research Group (University of Cork) to develop selfcare cards. These were found on evaluation to be written at a literacy level that was too high for patients to understand [43]. In a UK lifestyle change programme [46], volunteers were comprehensively involved in various aspects of the intervention including delivery, and gained confidence and skills. However, practical issues arose affecting the fidelity of the intervention including misaligned priorities between stakeholders. Similarly, one report described that too much information, often confusing, was put on a website developed with existing patients [45]. These practical issues highlight the need for PPI at the earliest level of research planning, which should be carefully monitored throughout, to ensure that PPI continues to be effective. These gaps and weaknesses should be remediated in future by the GRIPP2 guideline for development of tools to improve reporting of PPI, to understand the context, process, and impact of PPI for better conceptualization or theorization [51].

Strengths and limitations
This review represents a small subsample of a larger systematic survey of studies on health literacy interventions for older people. Potentially, using an existing inventory could result in studies being missed due to the inclusion criteria of that study not fully reflecting the purpose of the present review. For example, studies for the large inventory used for this review had to report outcomes, thus studies that may identify associations or other important issues around PPI were rejected [26]. Nevertheless, the studies were widely sourced and representative of the existing health literature for older people and the research raised concerns that could be addressed in a dedicated review on health literacy. We relied on screening abstracts at some stages which may have limited our samples; we do not pretend to have undertaken a thorough assessment exercise. We did not assess the intrinsic methodological quality of the studies included in the mapping exercise, neither were we able to compare studies reporting PPI with those that did not. We note that sometimes PPI representatives seemed to have blurred roles, such as also acting as research subjects [32,39], or co-investigators [38]. Blurring or dual roles is undesirable because it creates conflicts of interest; the PPI representatives may have cause and opportunity to bias outcomes. A more rigorous review than ours might exclude reports where roles appear to have been blurred.
Our choice of defining older adults as people aged 50+ is inevitably somewhat arbitrary. There is no universally agreed threshold for identifying relatively older adults. Age 50+ does reliably denote older adults as it is more than half of the average lifespan. For public health interventions, an "older adult" threshold at about 50 years may be desirable because this is a potential key window of opportunity to promote health literacy skills that might persist into retirement age, just before health literacy most declines and health care needs are likely to increase.
Production of this article itself did not directly involve patients or public representativesalthough consultations with such groups did inform the wider project that led to the creation of the larger inventory that was the starting point for our mapping review [18,26] . Within our selected studies, PPI was poorly reported. We checked for but failed to find PPI in earlier published developmental work, although we did not contact authors for more information. Many of the studies discussed here pre-date increasingly widespread obligations from funding bodies to orientate research design using PPI. Thus, this mapping review may be considered prescient or premature. Nevertheless, it is illuminating to map the gaps and deficiencies in existing research designs so that appropriate steps can be taken in future research and policy decisions.

Conclusions
Better reporting is required if the full potential valueand practical issuesof patient and public involvement in the research process are to be understood. Few of the health literacy studies in our review demonstrated patient and public involvement as integral to the research process adopted. Adding PPI to research can enhance quality and appropriateness of the research at every stage, but also holds cultural and ontological challenges for researchers used to being in control. For best effect, PPI should be included at the outset and continue throughout the research. High quality PPI can help develop the evidence base and inform practice for future interventions to improve health literacy among older people.