Patient feedback to improve quality of patient-centred care in public hospitals: a systematic review of the evidence

Background To review systematically the published literature relating to interventions informed by patient feedback for improvement to quality of care in hospital settings. Methods A systematic search was performed in the CINAHL, EMBASE, PsyInfo, MEDLINE, Cochrane Libraries, SCOPUS and Web of Science databases for English-language publications from January 2008 till October 2018 using a combination of MeSH-terms and keywords related to patient feedback, quality of health care, patient-centred care, program evaluation and public hospitals. The quality appraisal of the studies was conducted with the MMAT and the review protocol was published on PROSPERO. Narrative synthesis was used for evaluation of the effectiveness of the interventions on patient-centred quality of care. Results Twenty papers reporting 20 studies met the inclusion criteria, of these, there was one cluster RCT, three before and after studies, four cross-sectional studies and 12 organisational case studies. In the quality appraisal, 11 studies were rated low, five medium and only two of high methodological quality. Two studies could not be appraised because insufficient information was provided. The papers reported on interventions to improve communication with patients, professional practices in continuity of care and care transitions, responsiveness to patients, patient education, the physical hospital environment, use of patient feedback by staff and on quality improvement projects. However, quantitative outcomes were only provided for interventions in the areas of communication, professional practices in continuity of care and care transitions and responsiveness to patients. Multi-component interventions which targeted both individual and organisational levels were more effective than single interventions. Outcome measures reported in the studies were patient experiences across various diverse dimensions including, communication, responsiveness, coordination of and access to care, or patient satisfaction with waiting times, physical environment and staff courtesy. Conclusion Overall, it was found that there is limited evidence on the effectiveness of interventions, because few have been tested in well-designed trials, very few papers described the theoretical basis on which the intervention had been developed. Further research is needed to understand the choice and mechanism of action of the interventions used to improve patient experience.

(Continued from previous page) Conclusion: Overall, it was found that there is limited evidence on the effectiveness of interventions, because few have been tested in well-designed trials, very few papers described the theoretical basis on which the intervention had been developed. Further research is needed to understand the choice and mechanism of action of the interventions used to improve patient experience.
Keywords: Patient feedback, Patient experience, Patient-centred care, Quality improvement, Quality of care, Public hospitals Background Public health services have been moving towards putting patients at the centre of their care. Patient-centred care is defined as 'care that is respectful of and responsive to individual patient preferences, needs and values, and ensuring that the patient's values guide all clinical decisions [1]. Patient-centred care is considered to be one of the six domains of quality of care, where listening to and seeking to understand patients' perspectives of their needs, is key to the delivery of good quality care [1]. For greater clarity, the relationship between quality of care and patient-centred care is illustrated in Fig. 1.
This approach to care ( Fig. 1) promotes respect for patients' preferences and values, and provision of emotional support, physical comfort, information, communication and education, continuity and transition of care, coordination of care, access to care and the involvement of family and friends [2][3][4]. These have been shown to be associated with clinical benefits [5][6][7][8] and healthcare cost reductions [9][10][11].
Patient-centred care is assessed by patient feedback of their experience often referred to as patient experience measures [8]. It is becoming common for these measures to be collected routinely in order to monitor patientcentred care [12][13][14] . The U.S. and U.K., were among the first to develop and implement nationally standardised surveys for measuring patients' experiences. The American CAHPS (Consumer Assessment of Healthcare Providers and Systems) surveys were developed in 2008 and implemented in 2011. In the U.K., the reporting of the results of national standardised survey of NHS patients was made mandatory in their national health policy in 2010 [11,15]. Other countries such as Australia, Canada, Denmark, the Netherlands also established systems for collecting patient experience measures under their health policies, in their efforts to improve patient-centred care and other domains of quality of care [12][13][14].
Although the various methods of collecting patient experience such as complaints, compliments, surveys, interviews and focus groups have been widely researched Fig. 1 The relationship between quality of care and patient-centred care [16,17], there is still a debate about the use of the terms 'satisfaction' and 'experience' in these measures, which are sometimes used interchangeably [18,19]. Traditionally, up to the 1990s, patient satisfaction surveys were used to measure the quality of care from patients' perspective. However there were theoretical arguments against its sensitivity and usefulness, as 'satisfaction' was conceptualised as people's expectations and appraisals of care and not the actual 'experience' which relates to things that happened during care [19]. This led to the development of new patient experience surveys in the 2000s where the emphasis is on what happened to the patients during their hospital stay or clinic visit.
A recent systematic review [20], on the collection of patient experience and its use for quality improvements in health services, found that most quality improvement areas were in processes for admissions and producing patient education materials. Notably, these findings focused on areas that do not require changes to healthcare professionals' behaviour, yet many components of the patient experience are integral to the interactions, patients have with healthcare professionals.
Moreover, findings (results) from patient experience surveys frequently highlight the lack of time in consultations, difficulties in understanding tests and information from doctors and lack of details and specificity from the survey needed for quality improvements [21][22][23][24]. The lack of patient involvement in developing quality improvement initiatives, the insufficient expertise by healthcare professionals in conducting improvement work and lack of time and resources were some of the key barriers to efforts to improve quality of care [7,20,25].
Understanding which interventions are effective in improving the various dimensions of patient-centred care is needed to achieve good quality care. Improvement efforts in health services cannot be made without the feedback of patients, participation or changes on the part of the healthcare professionals and the resources and support of their organisations [26,27]. At present, it is unclear which interventions are effective and which behaviours need to change on the part of healthcare professionals and their organisations. The aim was to review the evidence about the impact of interventions informed by patient feedback on quality improvements in patientcentred care in hospital settings.

Search strategies
The research adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2009 checklist [28] and the review protocol was published (PROSPERO:CRD42018112964). The CINAHL, EMBASE, PsyInfo, MEDLINE, Cochrane Libraries, SCO-PUS and Web of Science electronic databases were searched. Search terms included a combination of keywords, MeSH-terms and text words related to feedback OR patient satisfaction OR patient preference AND quality of health care OR outcome and process assessment (healthcare) OR outcome assessment (healthcare) OR treatment outcome OR process assessment (healthcare) OR program evaluation OR quality assurance, health care OR quality improvement OR quality indicators, healthcare OR standard of care OR patient safety OR patient-centred care OR healthcare quality OR quality of service OR health outcome AND hospital, public were entered. The search was limited to published studies from January 2008 as the literature documented the development of patient experience surveys in U.S. in 2008 and the reporting of standardised patient experience survey results in other countries from 2010.

Study selection Inclusion and exclusion criteria
Studies were included if they had investigated an adult population, reported feedback from patients and quality improvements to care, published in an English peerreviewed journal from January 2008 till October 2018.
Using a standard form, information on study design, study setting, sample characteristics, sources of patient feedback, details of interventions used and outcomes were extracted by one author (EW) and verified by another author (JF). Where there was disagreement the third author (FM) reconciled the decision.

Assessment of study quality
The Mixed Methods Appraisal Tool (MMAT) [29] was used to assess study quality. The MMAT includes specific criteria for mixed methods studies, as well as for qualitative and quantitative studies. In MMAT revised (2018), the authors discouraged the use of an overall numerical score to reflect the quality of the studies but to provide a detailed presentation of the ratings of the criteria to reflect the quality of the included studies [30]. The assessment is made against five criteria, scored as 'Yes,' 'No' or 'Can't tell', and it was developed systematically [31]. For ease of discussion, in this review the studies were ranked as high (all criteria met), medium (four out of five criteria met) and low (three or less criteria met).

Data synthesis and analysis
Data synthesis allows researchers to critique and integrate research data from diverse disciplinary perspectives and studies which have used qualitative, quantitative, and mixed designs. Studies with multiple components intervention were coded to each of the intervention areas identified and according to the quality of the study; leading to some being counted more than once in the summary table. This approach is recommended for reviews seeking to understand the effectiveness of certain intervention areas, by categorising interventions by commonalities rather than considering the multiple components intervention as a whole unit [32,33]. Finally, a narrative synthesis was used to report the evaluation of the studies.

Search results
The initial search returned 1746 papers (Fig. 2), which were imported to Endnote and subsequently to Covidence [34] for screening; after removing duplicates, 1232 papers were retained. The title and abstracts were screened against the inclusion criteria. Two authors (EW and JF or EW and FM) assessed the papers and yielded 28 papers for inclusion. The final retention of 20 papers were made by consensus, any disagreements were resolved by consensus or consultation with a third author (FM or JF). The main reason for exclusion at this stage was that papers mentioned inclusion of patient feedback in the abstract but did not give any details of the patient feedback collected.

Quality assessment
Assessment of studies using MMAT indicated that two studies rated high fulfilling all five criteria, five studies rated medium meeting four criteria, 11 studies met only three or fewer criteria were rated as low and two studies could not be appraised because details regarding research aims, data collection methods or analysis were not provided (See Additional file 1).

Methodological characteristics and main findings
All studies included in the final review were based in hospital settings, and of these, three also included the health services' primary and community care settings as they formed part of the organisation. The methodological characteristics and main findings are summarised in Table 1.

Study location, sample and design
The studies were conducted in eight countries, UK (n = 6), Australia (n = 3), US (n = 3), the Netherlands (n = 2), Taiwan (n = 2), Canada (n = 1), Italy (n = 1), South Africa (n = 1) and the United Arab Emirates (n = 1). The different sources of feedback in the studies were interviews  Before-after study; Semistructured interviews were conducted with the participants, subsequently negative patient reviews on hospital rating sites and the hospital contextual details were emailed to the participants and they were interviewed again 4-6 months later.
10 Senior Health inspectors 23% of patient reviews were deemed relevant for risk identification by the senior health inspectors. The reviews which included major safety problems, severe damage or consequences for the patient and structural organisation problems, malfunction of doctor was deemed relevant.  (n = 7), patient experience surveys (n = 6), patients' narratives of their experience (n = 2), complaints (n = 2), patients' perception of service quality (n = 1), patient views on access (n = 1) and patient ratings online of hospitals (n = 1). A total of 77,588 participants contributed data to 17 studies, and participants characteristics were not reported in three [46,53,54]. The 20 studies were conducted in inpatient or outpatient settings in public hospitals with five studies providing additional details on the speciality settings. They included specialised cancer treatment (n = 3) and emergency medicine (n = 2).
The 20 studies comprised a cluster randomised control trial (n = 1), before and after studies (n = 3), crosssectional studies (n = 4), and organisational case studies (n = 12). The outcome measures in all the studies were on patient experience or patient satisfaction with waiting times, physical environment and courtesy of staff, which are components of the patient experience.  The study identified quality improvement priorities leading to changes in Emergency Departmentpalliative care processes. It also led to the creation of a patient-family-staff experience training DVD to encourage application of generic design principles for improving palliative care in the emergency department.

Areas of interventions
High clinicians' responsiveness to patients, patient education, the physical hospital environment, quality improvements, and improving the use of feedback are provided in Table 2. Only one of the studies [40] reported their theoretical basis and four studies [50,51,54,55] specified the use of quality improvement and experiencebased co-design methodology.

Communication
Interpersonal communications about health conditions and care transitions between patients and staff were the key area of intervention identified in improving patient experience in the studies in this review. The interventions targeted changes in staff's communication behaviours, for example, provision of weekly education sessions on communication skills and setting behavioural targets for staff. The frequency and mode of delivery of the education sessions reported were varied but they shared similar education content on customer service and interpersonal communication skills [36][37][38]. However, significant increase in satisfaction with explanation given and courtesy and efficiency of staff was only reported in Harnett et al.'s study [36] where the education component is part of a suite of other interventions.
In addition to staff education, two studies [37,38] also reported on organisational level interventions as part of the suite of intervention. Aboumater et al. [37] observed that hospitals with high patient experience scores promoted specific behaviours on communication and engagement of patients to staff using acronyms and slogans on (65%) and set standards and targets for staff for patient-centre and excellent service (60%). This observation is also noted by Buurman and colleagues [38] in their study where targets were set for staff, adoption rates of personalised communication with patients on discharge increased by 20% over 3 years. However, these changes cannot be assumed to be related to the interventions in the absence of a control group, in their study designs, it could be attributed to the passage of time or other factors.
Two further studies [51, 55] used experience-based codesign as an approach to engage, seek patient feedback on their experiences and views to identify improvements, discuss, design a suite of changes in communication, and professional practices. As the experience-based codesign methodology in its nature is about tailoring to the context, the findings from these studies may be limited to the experience of patients accessing cancer treatment services and emergency departments of hospitals. There was no measurement of patient experience, but the patients reported having had good experience when interviewed about the effects of the changes.

Professional practices in continuity of care and care transitions
Four studies highlighted discharge planning and associated care processes such as follow-up phone contact, Reduce repetitive assessments by multiple staff, plan for diagnosis giving in a dedicated space and provide written and verbal discharge information to patients, Responsiveness to patient (Respect for preferences and emotional support) [37,41,42,45,48,50] (n = 6) 1 5 Introduction of hourly proactive nursing rounds and weekly senior executive rounds, provide telephone contact to nurses regarding health concerns and clinical leads to review information flow about patient care along the care pathway.
The physical environment [43,50] (n = 2) 1 1 Made changes to floor coverings to reduce noise, creation of family rooms or quiet spaces in the hospitals Improve use of feedback [35,54] (n = 2) 1 1 Schedule meetings to discuss patient experience results and plan improvements and triangulate multiple sources of data to understand the feedback.
Quality improvement [43,48] (n = 3) 2 1 Provide structure and support by the organisation for the identification and implementation of quality improvements and monitor quality improvements as part of staff performance assessment.
a H high, M medium, L low, NA not accessed giving written discharge information to patients as a focus area in improving patient experience. It was found in two studies that use of both individual and organisational level interventions was significantly more likely to have a difference in patient experience. Aboumater and colleagues [37] reported that 52-56% of hospitals with high patient experience survey scores, indicative of high quality hospitalisation experience in their study, implemented multi-disciplinary rounds, follow-up with patients via phone calls post-discharge by nurses and used discharge folders for information sharing and consolidation. Organisational level interventions of using templates for personalised discharge letters, incorporating personalised discharge letters into the computer system of electronic medical records and integration of its use as hospital-wide policy were associated with an increase in the use of personalised discharge letter from 30 to 50% in the hospital over a 3 year period in Buurman et al.'s study [38]. Furthermore, two case studies [51,55] that provided an in-depth understanding of the experience-based co-design approach supports this observation between intervention to care processes and good patient experience. These studies explore the experience-based co-design approach in the redesign of palliative care and cancer care processes as part of a suite of interventions, where good experience was reported by interviewed patients.

Responsiveness to patient
The role of nurses was highlighted as a common component of the interventions employed in three studies [37,41,48], to improve the patient experience. The interventions targeted behaviours that were perceived by patients as respectful, courteous, prompt and personcentred. However, only weak associations between these interventions and positive patient experience were reported. In Abounmater et al. which used proactive nursing rounds (83%), and executives and leaders making rounds to engage and respond to patients (62%) [37], had high patient experience scores. Richard et al.'s cross-sectional study [41] observed that patients with nurse navigator support compared to those without reported higher satisfaction with waiting times. The role of doctors was generally not explored with the exception of Madden and Davis' study [42] where secondary data analysis was conducted to compare the results of two national patient experience surveys conducted in 2000 and 2004. It is interesting to note that this is the only study that reported a downward trend in aspects of patient experience with doctors (confidence in doctor and understanding of tests from doctors' explanation) for patients using breast cancer services in three health services. This was in spite of reported upward trend on a national level (across 172 health trusts in UK). The influences on this downtrend trend is unknown as there were no reported investigation on the probable causes or associations.

Patient education
Conceptualisation of patient education differed among studies. In Reeves and Seccombe's study [43], patients were given a comprehensive patient information pack about the discharge processes. This intervention was further complemented with the organisational level intervention of inclusion of its implementation action plans as part of staff performance assessment. While two other studies [42,50] did not provide details and defined it as information for patients. There was no significant evidence on any association or efficacy of interventions in this area from these studies.

The physical environment
Interventions to improve the physical environment found in two studies, focused on engaging patients in the redesign of physical spaces in the emergency department [50] and reduction of noise levels in the hospital [43]. Overall, the changes in the physical environment could not be solely associated with changes to the patient experience, as these interventions were part of a larger suite of interventions.

Improve use of feedback
Reeves and West's study was the only cluster RCT [35] in this review. They found significantly better experience survey scores among patients in the condition where wards held facilitated meetings to review patient feedback and plan improvements compared to the two other conditions (feedback sent to the Matron of ward and feedback on ward level sent to individual nurses). From the findings of the study, the authors hypothesised that it is the opportunity for facilitated delivery of the feedback to nurses that increased the acceptability of the feedback that prompted the change in behaviour.

Quality improvement
The studies [43,48] that investigated interventions used in quality improvement projects suggested that it is necessary to approach this at both the staff and organisational levels. They observed that good patient experience was possible when there was regular data review, effective implementation of action plans, and incentives and staff performance review by their organisations.

Discussion
The results of this review show that interventions employed in the included studies, predominantly target and support the theoretical dimensions of patientcentred care. Interpersonal communication between healthcare professionals and patients about their health conditions and care, processes affecting care continuity and discharge planning and showing respect for patient preferences and providing emotional support clearly emerged as important intervention areas, most frequently noted in the 20 studies. However, the efficacy of the interventions must be interpreted with caution because causal relationships were mostly not tested in the studies included in this review.

Strengths and limitations
The strength of this review is the specific focus and inclusion of the use of patient feedback for improving patient-centred care in the search strategy for the review. The search strategy was designed in consultation with an information analyst, to produce a replicable search for all relevant multiple databases, using MeSH search terms and the inclusion of all study designs, single and multiple interventions and variety of countries, to provide a search of the evidence that has been applied to the existing context in health services rather than just research settings.
We acknowledged some limitations in this review, only studies published in English language and after January 2008 were included. There could be other relevant studies published prior and in other languages that were missed. Further details on the interventions in the included studies could also be missed as no further contact was made with their respective authors.

Main findings
There are several possible explanations for this weak body of evidence on the efficacy of the various interventions, firstly, the study designs employed in the studies were mainly correlational and qualitative and secondly, the quality of the studies. There is only one cluster RCT in this review that provided evidence that patient feedback was effective in improving quality of care when it was facilitated and discussed with nurses and planned for at ward level compared to other conditions where it was not facilitated or discussed. Overall, 11 studies reported improvement in patient experience outcomes, but only five studies quantified their findings by reporting on the changes in outcome measures.
The quality of evidence of the five quantitative studies that reported outcome measures was low, beyond the limited representativeness of the study populations in some of the studies, the weak associations between the interventions and outcomes with no acknowledgment of potential confounders such as the passage of time.
The qualitative studies in this review were more varied in study quality, four of the studies were conducted well with detailed reporting. The studies highlighted how experience-based co-design methodology was utilised in acute care settings to engage and partner patients in making improvements to care and also contribute to the understanding of the areas of care that were deemed important by patients.
Studies that used multiple interventions targeting change on both individual and organisational levels were associated with better outcomes than those studies with single interventions. This review found that interpersonal communication training for healthcare professionals combined with organisational policies of setting targets and promoting behavioural standards for the staff were associated with improved (increased) patient experience. Similarly, this association was also found with implementing processes and practices with multidisciplinary team meetings and sharing of discharge information practices, in conjunction with organisational policies of setting targets and promoting behavioural standards for the staff.
These findings are in line with studies [56,57] that explored a system view in implementing interventions where considerations are given to mediating factors organised by structure (organisational characteristics), process (care processes) and outcome (patient experience, clinical outcomes) using Donabedian's model. This is further supported by findings from studies [58,59] investigating factors needed for successful implementation and integration of interventions to routine work using the Normalisation Process Theory [60,61]. With the acknowledgement of targeting change on multiple levels within a system, it is not surprising that there is a growing body of literature on developing and evaluating multiple components interventions [62].
Beyond the limitations of the study designs and quality of the studies, a plausible explanation for the weak evidence is the lack of explicit use of theory in the intervention development or discussion of results in the majority of the studies. The importance of using theory is reflected in the growing research of using behavioural and organisational theories in the design of interventions involving professional practice and the understanding of which mechanism or elements of the interventions are the most important [63]. For example, in the studies targeting improving communications between patients and healthcare professionals, the effective interventions were using a combination of educational sessions for staff and action planning and monitoring interventions by organisations. Without being explicit about their theory of change, in the communication behaviours in those studies, it is plausible that educational sessions for staff were conducted to engage staff on communication as a priority, instead of their lack of skills. If that was true, more targeted interventions to address engagement and prioritisation by healthcare professionals could be more effective.
There are different theories that may be relevant for developing interventions at multiple levels, using approaches that address, cognitive, educational and organisational theories that can contribute to changing healthcare professionals' behaviours [64]. For example, theories such as theory of planned behaviour and social learning theory [65][66][67] may be more relevant to interventions directed at individuals and teams. On the other hand, organisational theories such as Continuous Quality Improvement [68] and organisational quality culture [69] may be more relevant to interventions directed at service redesign for the whole hospital with multiple stakeholders [63].

Further research
From the review findings, the field of research could explore the gap in the explicit use of theory in their target for change and choice of interventions. This will enable the comparison of interventions and their mechanism of action, across settings to build the evidence base. Beyond those interventions found in this review, another gap to address is the lack of research in the interventions targeting the emotional support, access to care, involvement of family and friends dimensions of patientcentred care. It could provide further insights in the interpersonal relationship between patients, their family and the healthcare professional and its impact on patient-centred care.
There is also room for further progress in examining the acceptance and utilisation of patient experience in the development and evaluation of improvement efforts in patient-centred care. Despite the widely acknowledged concept of patient-centred care, the low number of studies found in this review that includes patients' perspective and experiences of care is professionally and practically concerning.
The conceptual definitions and differentiation between patient experience and satisfaction are still debated in the existing literature [19,70,71]. However, in the reviewed studies the authors did not differentiate between these concepts. In some studies in the measurement of patient satisfaction, the focus was on the experience of the process and feeling, rather than the concept of satisfaction where their expectations are met or not. In other studies on patient experience measurement, the focus was on the patients' expectations. The lack of conceptual differentiation of these concepts could be addressed in future studies as there are implications in their operationalisation and comparability of findings.

Implications for practice
The evidence from the reviewed studies suggests that health services are collecting feedback from patients on their experience either locally or through nationally standardised survey instruments and increasingly reporting them as one of their performance indicators. Not surprisingly, the collection and reporting of patient experience in itself, does not improve care. Considering the evidence from the review, the patient experience collected needs to be discussed and facilitated with healthcare professionals in their respective operational units in order to provide opportunities for them to engage and act on the feedback to improve care.
The finding on the strong focus on interventions targeting communication between healthcare professionals and patients suggests that communication is akin to the 'delivery' system for the dimensions of patient-centred care. This could be a consideration for health services as a starting point as it has also been recommended as an area of focus with good cost-benefit to health services [72].

Conclusion
This review shows that incorporating patient feedback of their experience into research on quality patient-centred care is still an emerging field. The limitations outlined show that the degree of effectiveness attached to the different interventions must be interpreted with caution. However, the findings of this review can inform researchers, healthcare professionals, health systems and policy makers to focus on interventions, practice guidelines and strategies that incorporates patient feedback of their experience in patient-centred care improvement work. Care is truly patient-centred when it is guided by the perspective of the one that matters -the patient.
Additional file 1. Risk of bias assessment. Authors' contributions EW contributed to the concept, data collection, data analysis, data interpretation and drafting the manuscript. FM and JF contributed to the concept, data analysis and manuscript review. All authors reviewed and approved the final version of the manuscript.

Funding
Sources of support: Eunice Wong is supported by a research higher degree scholarship jointly funded by the Australian Government Research Training Program and Safer Care Victoria. Jane Fisher is supported by the Finkel Professional Fellowship, funded by the Finkel Family Foundation. Beyond funding support, the funders had no direct role in the study design, data collection, analysis, or interpretation, or writing the manuscript.