The impact of health literacy environment on patient stress: a systematic review

Background There exists little literature on situational health literacy - that is, how an individual’s health literacy varies across different health literacy environments. However, one can consider the role of stress when examining the relationship between health situations and decision-making ability, and by proxy health literacy. The aim of this study was to assess the strength of the evidence on the relationship between health situations and patient stress, considered in the context of health professional perception, and determine what health situations act to influence patient stress. Methods A systematic review of English articles using PubMed, PsycINFO, CINAHL and Embase databases was conducted. Search terms focused on ‘patient’, ‘stress’, and ‘health care situations’. Only peer-reviewed original research with data on patient stress in the context of a health facility environment was included. Studies were screened and critically appraised by both authors. Study elements for extraction were defined by RO and extracted by JY. Results Twenty-four studies were included for narrative synthesis. Patients in Intensive Care Units were more stressed about factors relating to their physical discomfort, with some agreement from health care professionals. Parents of children in Intensive Care Units were more concerned with stressors relating to their child’s appearance and behaviour, and alteration in their parental role. Few studies examined health settings other than Intensive Care Units, and those that did varied greatly in terms of study design and population characteristics, lacking generalisability. Conclusions Overall, the findings of what patients find most stressful in Intensive Care Units can guide health care professionals practicing best practice care. However, the evidence on how patient stress is influenced by non-Intensive Care Unit health care settings is weak. Further research is needed to enhance current understanding of the interaction between patient stress and health care environments in both hospital and primary care settings.


Health literacy
Health literacy can be defined as "the degree to which individuals have the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions" [1]. Its importance is well established. Not only is low health literacy associated with poor health for individuals and poor financial outcomes for health systems [2,3], it is also highly prevalent both in Australia and internationally [1,4].
Increasingly, health literacy is recognised as dynamic, varying depending on the context and characteristics of individuals, environmental and social factors, and the demands and burdens placed on individuals [1,[5][6][7][8][9][10][11][12]. Indeed, a critical component of health literacy is the health literacy environment, which is the "infrastructure, policies, processes, materials, people and relationships that make up the health system and have an impact on the way that people access, understand, appraise and apply health-related information and services" [13]. However, despite a few key studies [14,15], there is still little current literature on situational health literacythat is, how an individual's health literacy varies across different health literacy environments.
Stress can be considered when examining the relationship between health situations and decision-making ability, and by proxy health literacy. Much of the literature suggests that stress impairs decision-making ability [16,17], through several mechanisms. Firstly, high stress levels have been found to cause inadequate consideration of alternatives, resulting in dysfunctional strategy use [16]. Additionally, there is an over-reliance on intuitive decision-making rather than strategic choice, reflecting an insufficient adjustment from automated responses [16]. Furthermore, stress triggers altered feedback processing, causing stressed individuals to make more disadvantageous choices [16]. Given the clear influence of stress on decision-making ability [16,17], the focus of the current study is on what health situations act to influence patient stress.

Health situations and stress
Environmental factors are known to play a large role in increasing or decreasing patient stress [18]. Much of the literature focuses on stressors, both physical and psychological, associated with hospitalisation. These include lack of natural light, increased noise levels, presence of unwelcome smells or other sensations, absence of clocks, and a perception of crowding by unfamiliar people [19][20][21]. Intensive Care Units (ICUs) especially are known to be stressful environments [19], with specific stressors including experiencing pain, disruption of sleep-wake patterns, and intubation [22,23]. A perceived lack of control over these physical environmental factors also increases patient stress [21,24]. In addition, hospital patients are stressed by their inability to obtain desired information, and their fear of an unknown or serious illness [24]. Furthermore, decreased social support, such as with loss of contact with family and friends, and financial worries associated with hospitalisation and illness, have been identified as stressful for patients [19,22,24].
Stress-reducing environmental factors have also been identified. Viewing plants and nature has been shown to decrease patient stress [25], an effect which was reproduced by viewing photographs of plants or nature in a hospital waiting room [26]. In addition, environmental interventions that counter stress-inducing factors are thought to reduce patient stress [18]. Recommendations have been made for single-bed rather than multi-bed rooms, and elimination of noise sources, as these would lead to improved sleep [18]. Similarly, administrative and procedural information, external building cues, local information systems, and global structural redesign, have also been recommended as these interventions reduce spatial disorientation, and consequently decrease patient stress [18,27]. These recommendations appear to also apply beyond hospitals to the primary care setting [28].
There are also other aspects of health situations that influence patient stress. These include the providerconsumer interactionthat is, what physicians say and how they deliver this information [29]. In particular, taking time, empathising with patients, and overall effective communication skills have been shown to help reduce patient stress [29,30]. Additionally, the broader social, economic, and psychosocial contexts of stress can be considered. Constant exposure to socially and economically challenging environments has been shown to increase stress [31], while strong social support networks protect against stress and other environmental threats to health [31,32].
Clearly, many health situations affect patient stress, potentially influencing decision-making ability and by proxy health literacy. However, the cumulative strength of the evidence on this relationship between health situations and patient stress has not been evaluated. The aim of this systematic review is to assess the strength of this evidence, and determine what health situations act to influence patient stress. In doing so, this study will also inform our understanding of health situations in supporting or constraining health literacy.

Methods
The study design of this systematic review was guided by the PRISMA Statement [33] (Appendix 1). Studies were identified by searching through the electronic databases PubMed, PsycINFO, CINAHL and Embase, with coverage from 1997 to February 2017. This timeframe was selected to provide a long enough period to comprehensively review the research, whilst maintaining the relevance of the health care context to contemporary practice. No grey literature was reviewed, and the reference lists of included articles were not searched for additional studies.
The search terms comprised of three components: patient; stress; and health care environment. Index terms were used when possible as appropriate for each database, and synonyms were included. Terms relating to stress of health care professionals or staff, and mental disorders of patients, were excluded from the search. In addition, the search was limited to journal articles published in English since 1997. The full search strategy is provided in Appendix 2.
The study selection process is summarised in Fig. 1. Articles identified through the search process were screened by JY based on titles alone, removing articles that were not relevant to the current study question. The abstracts of the remaining articles were then assessed independently by JY and RO, applying the predetermined inclusion and exclusion criteria (Table 1). Articles without abstracts were removed at this point. Disagreements regarding article inclusion or exclusion were resolved by a third reviewer, NA. Articles that were evaluated as sufficiently rigorous were included for critical appraisal. This process was completed independently by JY and RO, using an appraisal template adapted from Bush et al. [34] (Additional file 1). Articles of lower methodological strength, generally corresponding to a Fig. 1 Flow diagram of study selection process, including formal search, screening, application of inclusion and exclusion criteria, and critical analysis, with the number of articles included and excluded at each step. Adapted from the PRISMA Statement [33]  score of 0 or failing to score 2 in any of Appraisal Questions 1-5, were excluded. Discrepancies regarding the eligibility of articles at this stage were resolved by discussion between the two reviewers until agreement was reached. Eligible articles were included in the qualitative analysis. Data extraction items were defined by RO and extracted by JY from eligible articles. These data items included: study source, design, country, population characteristics, health care setting, measured variables, intervention, and outcomes. Due to the nature of the types of studies included, a narrative synthesis, rather than meta-analysis, was the more appropriate method of analysis.

Results
A total of 1506 unique articles were identified through the database searches conducted. Of these, 874 articles were excluded based on title-scanning as they were not addressing the research question. A further 573 were excluded based on abstract-scanning as they did not meet the inclusion and exclusion criteria. Full texts of the remaining 59 articles were critically appraised, and 24 were further excluded based on exclusion criteria whilst 11 were excluded after formal appraisal of methodological strength. The included articles received appraisal scores ranging from six to nine, out of a possible twelve. Overall, 24 articles were assessed to be eligible for data extraction and narrative synthesis, and were included in this systematic review (Fig. 1).
The study designs of included research were mostly descriptiveof the 24 studies, nineteen (79%) were cross-sectional surveys, one used a descriptive qualitative design, and only four were experimental. Included studies were also internationally diverse. Nine studies were set in Asia, eight in Europe, six in the USA, two in Brazil, and one in New Zealand. Furthermore, all studies were conducted in hospital settings, with the majority (71%) set in ICUs (nine adult, eight paediatric/neonatal). Regarding participants, eight studies focused on parents and sixteen on patients, with five of the latter also including health care professionals. Most participants were adults, with only Causey et al. [35] recruiting children and adolescents, and Yeh et al. [36] recruiting patients as young as fifteen years. Finally, the measured variable, stress, was assessed with validated tools in most studies (92%), while Larsen, Larsen and Birkelund [37] conducted participant observation and interviews, and Lilja, Ryden and Fridlund [38] measured serum cortisol ( Table 2).

Patient and parental stressors in ICU
Considering the nine studies that ranked patient stressors in ICU [19,22,23,[39][40][41][42][43][44], the items from the ICU Stressor (or similar) scales that were rated highly most frequently related to physical discomfort. As examples, top ten stressors included 'being in pain' in all studies, and 'being unable to sleep' in all but one study. Overall, seven of thirteen (54%) items in the physical discomfort subscale were listed as top ten in two or more studies. In contrast, only 35% of stressors relating to psychological distress, and 36% stressors relating to the ICU environment, were ranked highly in at least two studies. No stressors relating to treatment procedures were listed more than once in the top ten (Table 3).
Comparing stressors as ranked by patients with health care professional rankings, shows that some stressors were also ranked consistently highly by health care professionals across studies, such as 'being in pain', 'being tied down by tubes', and 'having tubes in your nose/ mouth'. However, for the most part, stressors were evaluated differently by patients and health care professionals, with significant differences between scores on stressor scales in some studies [22,23]. In particular, Biancofiore et al. [22] found that patients used a score of 1 (not stressful) more frequently than health care professionals when rating stressors, while health care professionals tended to use scores of 2, 3 and 4 (increasing stressfulness) more frequently (Table 3).
For parents, stressors relating to the appearance and behaviour of their child, or alteration of their parental role, were ranked most highly in almost all studies. Parental role alteration, and child appearance and behaviour were ranked first and second respectively by parents in studies of Franck et al. [47], Nizam and Norzila [51], and Reid and Bramwell. On the other hand, stressors relating to staff communication and behaviour, and the ICU environment, were consistently at the bottom of parent rankings (Table 3).

Non-ICU settings
Seven of the 24 studies examined health care environments other than ICU. These included a psychiatric inpatient unit; medical, oncology and surgical wards; haemodialysis units across various medical centres, regional and community hospitals; and a radiology department waiting room. The designs of these seven studies varied greatly, ranging from cross-sectional survey, to qualitative and experimental, the interventions of which showed inconsistent effects ( Table 2).
Children and adolescents admitted to an inpatient psychiatric unit appeared to be more stressed by separation from their friends and family, and their loss of autonomy, than by other aspects of the health care environment such as the psychiatric setting and staff interaction [35]. More broadly, stressors were found to be perceived differently by patients depending on the type of health facility. Yeh et al. [36] showed that issues such as Larsen, Larsen and Birkelund [37] found, through interviews with cancer patients, several recurring themes relating to the health care environment. These included the health effects, both positive and negative, of a health care setting, and the implications of (lack of) patient privacy, such as withholding information from doctors, avoiding   6, 3), 2) being in pain (3, 2, 1), 3) having tubes in nose/mouth (3, 2, 1), 4) missing husband/wife (5,9,9), 5) seeing family & friends only a few minutes a day (1, 7, 11), 6) being tied down by tubes (7,8,6), 7) being thirsty (6,11,4), 8) hearing the heart alarm (11,12,8), 9) having no control over oneself (8,5,5) (1), 2) not being in control of yourself (9), 3) not being able to sleep (11), 4) hearing the buzzers and alarms from the machinery (4), 5) being thirsty (40), 6) being in pain (6), 7) not knowing when to expect things will be done to you (8) other patients, and preferring single-bed rooms. Despite these apparent shortcomings of their accommodation, the patients maintained a sense of acceptance of and resignation to the hospital environment. Several studies investigated whether altering aspects of the health care setting would impact patient stress. Beukeboom, Langeveld and Tanja-Dijkstra [26] tested the effect of exposure to nature on stress, and found that both real plants and posters of plants significantly lowered the stress of patients in a waiting room. This effect was at least partially mediated by the perceived attractiveness of the room. Additionally, music therapy appeared to decrease both psychosocial and physiological stress for patients in haemodialysis units [53]. On the other hand, preoperative information provided by an anaesthetic nurse had no significant impact on patient stress [38]. Only one study compared inpatient and outpatient settings, which found no significant differences in stress between the two groups, in patients awaiting pacemaker implantation [54].

Summary of evidence
There are several main findings regarding what was identified as stressful about health care settings. Firstly, patients in ICU were consistently more stressed about factors relating to their physical discomfort, than those relating to psychological distress or the ICU environment. Meanwhile, apart from some similarity with stressors relating to physical discomfort, health care professionals placed importance on different stressors to patients, and generally rated them as more highly stressful to patients than did patients. Additionally, parents of children in ICU consistently placed the most importance on stressors relating to their child's appearance and behaviour, and alterations in their parental role, while being less concerned about stressors relating to staff communication and behaviour, and the sights and sounds of the ICU environment. Overall, the strength of the evidence for these findings is acceptable, and while difficult to apply to non-ICU settings, can be expected to be consistent between ICUs especially given the international diversity of the included studies.
Adult patients in an oncology ward often found their lack of privacy, related to the architecture in multiple-bed rooms, stressful and non-healing [37]. That this finding is not entirely consistent with the above studies, where stressors relating to the health care environment were less worrisome, may be explained by the different study designs In both inpatients & outpatients, subjective stress decreased from pre-op, to day 1, to day 3/4 -no significant differences in stress between groups at any time ICU intensive care unit; PICU paediatric intensive care unit; GCU general care unit; HCP health care professional; SD standard deviation and health settings. In addition, children and adolescents in a psychiatric unit appeared to find separation from family and friends, and loss of autonomy, more stressful than staff interactions and the psychiatric environment. However, the evidence for this finding is less robust given no studies other than that of Causey et al. [35] investigated this population and health care setting. The nature of the health care setting appeared to have variable effect on patient stress. Broadly, the type of health care facility, specifically its ownership, influenced how stressors were perceived by its patients [36]. In contrast, no differences in levels of patient stress were found between inpatient and outpatient health settings [54]. Admittedly, these results are difficult to compare as the studies were very different, in terms of study design, population characteristics, and assessment tool.
The experimental studies found that exposure to nature and plants, and music therapy, decreased patient stress, while pre-operative provision of information had no such reduction on patient stress. These findings, while promising, carry less weight than the previous outcomes. One of the main reasons for this is that across the few included experimental studies, the population, setting, interventions, and measured outcomes all differ. Furthermore, they lack generalizabilityfor example, the music intervention used by Cantekin and Tan [53] was Turkish art music songs, the effect of which may be difficult to reproduce in other populations. Therefore, the evidence of these outcomes is relatively weak.
This systematic review set out to firstly assess the strength of evidence on the relationship between health care situations and patient stress, and secondly determine what health care situations act to influence patient stress. It has partially met these aims. Certainly, this study confirms that ICUs are indeed stressful environments and produces some detail on specific stressors; however, the evidence for other outcomes in different health care settings is insufficiently robust.
The importance of stress in the relationship between health situations and decision-making ability relates to the way high stress levels result in dysfunctional strategies, over-reliance on intuitive decision-making, and more disadvantageous choices [16]. This implies that patients in stressful health situations such as ICU would have impaired decision-making abilities and likely constrained health literacy. How specific features of these health situations lead to patient stress, and how stressors in these contexts affect decision-making ability are important areas to examine further.

Limitations
The study has several limitations. Firstly, the review process may have been limited by the search strategy. The diversity of the literature around stress and healthcare made it difficult to develop a strategy with sufficient sensitivity and specificity. This meant that a large variety of articles, often tangential to the topic, were identified, and additionally the risk of missing key articles was increased. This may have been exacerbated by the restriction to only English-language publications. Furthermore, no additional searches were conducted for grey literature, and articles were not drawn from reference lists of included studies, or those citing included studies. This may have introduced a degree of publication bias. Another limitation was the lack of weighted scoring in the appraisal tool, meaning no explicit cut-off value could be set as an inclusion or exclusion criterion.
In addition, a limitation of the data analysis was the diverse nature of the included studies, with study designs, patient populations, health settings, interventions, and outcome measure often differing across studies. Moreover, a narrative synthesis, which was necessary given the types of include studies, cannot provide the same degree of strength of evidence as a meta-analysis. That the main findings of the analysis were drawn mostly from studies set in the ICU also limits the generalizability of the review. One issue specific to these studies is that using the stressor scales as assessment tools presupposes the factors that patients find stressful, and may not account for other relevant stressors.

Implications
An important next step would be to investigate patient stress in the context of health care settings other than the ICU. These should include not just other aspects of the hospital setting, such as different wards or waiting rooms, but also extend to the primary care setting. While an enhanced primary care environment has been associated with decreased patient anxiety and improved patient satisfaction [28], further research could enhance current understanding of the interaction between patient stress and this environment. As individual stressors are increasingly identified, future research could also investigate the effects of interventions that target these stressors. Ultimately, given the original focus on decision-making and health literacy, future research must directly investigate health literacy levels of patients and how this is influenced by different health care settings. Furthermore, this research should examine other aspects of the health literacy environment, such as relationships, infrastructure, and policies. Studies of this kind will more likely have lasting impact on clinical practice and health care design.

Conclusions
Overall, this systematic review has revealed that the evidence on how patient stress is influenced by non-ICU health care settings is weak. That being said, this review does show what patients find most stressful in the ICU, and these findings can guide health care professionals practicing best practice care. 15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies).

Additional analyses
16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified.

Study selection
17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram.

8
Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations. 8,9,16,17 Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12).

17-20
Results of individual studies 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot.

Summary of evidence
24 Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers).

11-13
Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias).

13, 14
Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research.

14, 15
FUNDING Funding 27 Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review. For more information, visit: www.prisma-statement. org.
Page 2 of 2

APPENDIX 2
Search strategy: PubMed