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

Objective: 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, 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.

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 (18). 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 (19), 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 areas 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 (20)(21)(22)(23). 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 (7,19,24).
Understanding which interventions are effective in improving the various dimensions of patientcentred 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. 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 patient-centred care in hospital settings.

Method Search strategies
The research adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2009 checklist (25) and the review protocol was published (PROSPERO:CRD42018112964).
The CINAHL, EMBASE, PsyINFO, MEDLINE, Cochrane Libraries, SCOPUS 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 AND hospital, public were entered. The search was limited to published studies from

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 peer-reviewed 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).

Assessment of study quality
The Mixed Methods Appraisal Tool (MMAT) (26) 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 (27). The assessment is made against five criteria, scored as 'Yes,' 'No' or 'Can't tell', and it was developed systematically (28). 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 component interventions 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. 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 (29) for screening; after removing duplicates, 1232 papers were retained. The title and abstracts were screened against the inclusion criteria. This initial screening yielded 28 papers after which the copies of the full-text were retrieved and assessed by two authors for inclusion.
Disagreements were resolved by consensus or consultation with a third author (FM or JF), of which 20 papers were retained for final inclusion. 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.
A total of 77,588 participants contributed data to 17 studies, and participants characteristics were not reported in three (30)(31)(32). Specialised cancer treatment centres (n = 3), emergency departments of hospitals (n = 2) and dedicated veteran hospital (n = 1) were the target settings in six studies, the rest of the studies were inpatient or outpatient settings in public hospitals.
The 20 studies comprised a cluster randomised control trial (n = 1), before and after studies (n = 3), cross-sectional 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.

Intervention Areas of interventions
The interventions proposed and implemented in the studies were in the areas of communication with patients, professional practices, clinicians' responsiveness to patients, patient education, the physical hospital environment, quality improvements, and improving the use of feedback (Table 2). Only one of the studies (33) reported their theoretical basis and four studies (30,(34)(35)(36) specified the use of quality improvement and experience-based co-design methodology.  In addition to staff education, two studies (38,39) also reported on organisational level interventions as part of the suite of intervention. Aboumater et al. (39) 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 three 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 (34,36) used experience-based co-design 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 co-design 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, 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 (49) 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 30-50% in the hospital over a three year period in Buurman et al.'s study (38). Furthermore, two case studies (34,36) 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 experiencebased 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 (44,45,49), to improve the patient experience. The interventions targeted behaviours that were perceived by patients as respectful, courteous, prompt and person-centred. However, only weak associations were reported with patient experience. In Abounmater et al. which used proactive nursing rounds (83%), and executives and leaders making rounds to engage and respond to patients (62%) (49), had high patient experience scores,. Richard et al.'s cross-sectional study (44) 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 (40) 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 (47), 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 (35,40) 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 (35) and reduction of noise levels in the hospital (47). 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 (48) 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 (45,47) 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.

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, despite handsearching reference lists of papers, did not yield additional relevant studies.

Main findings
No causal relationships between interventions in the dimensions of patient-centred care and patient experience outcomes was established in the studies due to the study designs employed. 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, can only provide some indication that the interventions could be associated with changes in patient experience outcomes.
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. Unfortunately, neither the effectiveness of each intervention nor interaction effects of the interventions were investigated in these studies, and it is unclear which interventions or combination of interventions contributed most to improving patient experience.
Nevertheless, there are some indications that some of the interventions are likely to have some effect on patient experience outcomes. Interpersonal communication training for healthcare professionals combined with organisational policies of setting targets and promoting behavioural standards for the staff were indicative of increased patient experience. Processes and practices in regards to the use of multidisciplinary team meetings and sharing of discharge information practices, in conjunction with organisational policies of setting targets and promoting behavioural standards for the staff.
Beyond the limitations of the study designs and quality of the studies, a plausible explanation for the variability in the effectiveness of the interventions and weak evidence is the lack of a theoretical basis 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 (50). For example, in the studies targeting improving communications between patients and healthcare professionals, the effective interventions were using a combination of education 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 education sessions for staff was conducted to engage staff on communication as a priority, instead their lack of skills. If that was true, more targeted interventions to address engagement and prioritisation by healthcare professionals could be more effective.

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 patient-centred care.
The conceptual definitions and differentiation between patient experience and satisfaction are still debated in the existing literature (18). 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 and conversely in some studies on patient experience measurement, the focus was on the patients' expectations. This could be addressed in future studies.
The evidence and quality of the studies in the review could be an indication that the use of patient feedback for improvement to the quality of care is still emerging. Research on effective implementation of interventions highlighted the importance of context (setting, organisational resources, leadership). This review supports the role, organisations play, when it complements interventions to change health professionals' behaviours.

Ethics approval and consent to participate
Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analysed during this study are included in this published article and its supplementary information file.

Competing interests
The authors declare that they have no competing interests

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.

Authors' contribution
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.