Patients experiences. Top heavy with research. Sarah Russell. Literature Review June

Transcription

1 Patients experiences Top heavy with research Sarah Russe Literature Review June

2 Patients experiences: Top heavy with research Literature Review 2013 Correspondence v Research Matters Sarah Russe PhD, BA (Hons) Principa Researcher PO Box 1235 Fitzroy North VIC 3070 Emai: sarahrusse@comcen.com.au Website: Design and desktop pubishing: MacNificent Cover iustration: Top heavy with research by Erica Evans Bayside Medicare Loca A rights reserved. Russe S (2013) Patients Experiences: Top heavy with research Research Matters: Mebourne Acknowedgements Our thanks to Jason Wasiak for his skis and expertise in searching academic databases, and Erica Evans for once again capturing a compex idea in a drawing.

3 Tabe of Contents Summary Introduction Literature search Patient feedback Terminoogy Types of patient feedback... 3 Key messages from Section Coecting patients feedback Samping Timing of data coection Methods... 7 Key messages from Section Survey instruments Key Messages from Section Key variabes that determine patients experiences Patient characteristics Organisationa characteristics Key messages from Section Patients Experiences: Top Heavy with Research iii

4 6 Purpose of patients feedback Provide information Measure quaity of heath care services Improve quaity of heath care Change professiona practice Evauate innovations Improve the quaity of specific interventions Compare heath services Measure coordination of care Rate heath care services Improve compiance Key messages from Section What we know about patients experiences of primary care What studies te us Disseminating resuts of studies Key messages from Section Future for patients experiences Patient Experience Framework Key messages from Section Concusion References iv Patients Experiences: Top Heavy with Research

5 Summary Consumer advocates have argued for many years about the need to improve the quaity of heath care from service users perspectives. Prior to 1995, research on patients experiences was sma-scae and reied mosty on quaitative methods. However, once this type of research became mainstream, the sampe sizes became arge and the methods mosty quantitative. The UK s annua GP Patient Survey, for exampe, incudes over 5.5 miion peope. During the past three years aone, over 2,100 peerreviewed artices were pubished on patient reported outcomes, mosty patients satisfaction and experiences. These artices are remarkabe for their repetitiveness. Athough studies focus on different sites of heath care or on a specific iness and use different methods and various instruments the existing studies draw simiar concusions. Most patients are satisfied with the heath care they receive. Even those patients who have bad experiences are generay satisfied with their heath care. Heath care organisations spend a considerabe amount of time and resources on gathering data on patients feedback. Most studies focus on a specific iness or sector in the heath care system. This burgeoning interest in patient feedback refects a shift towards patient-centred care. However, strategies for coection, coation, anaysis and dissemination of patients experiences remain ad hoc. In addition, a number of different instruments are used to describe and measure patients experiences. Without standardised surveys, it is impossibe to compare findings with other heath services, or often even within the same service over time. Patient satisfaction surveys remain the most common type of feedback though without a universa definition of satisfaction, measurements of patients satisfaction are probematic. In addition, findings from satisfaction surveys are non-specific, making them useess for improving patients experiences. Patients experiences provide a more discriminating measure of a heath service s quaity than questions about satisfaction. However, reativey minor aspects of a heath care consutation may have a significant impact on patients experiences (but not on their cinica outcome). Evidence aso indicates that patients experiences are infuenced by socio-demographic factors. This raises an interesting question: Does this refect different expectations among different types of patients? Or do different types of patients within the same heath service receive different types of care? The focus to date has been on coecting data on patients experiences rather than using the findings to improve service quaity. In fact, itte is known about how such feedback can be used to improve patient-centered care. There is some evidence that data coected at the eve of individua teams, and cose to the time when the care was experienced, may have the greatest impact on services. A number of different methods have been used to measure patients experiences, dividing broady into quaitative and quantitative methodoogies. Both methodoogies are usefu though for different purposes. The key to effective data coection is to use mutipe methods and a range of data sources (incuding socia media such as bogs, Twitter, Facebook, and rating websites). Mutipe methods wi enhance representation and therefore the vaidity of research findings. The iterature highights individua, organisationa and systemic barriers to using patients feedback. One important barrier is professiona scepticism about its vaue. Some practitioners argue that patients are not medica experts, and their perspective is therefore of no vaue. Heath care practitioners may be experts about medica treatments, but patients are experts about their own ives. Patients ceary have the capacity to report on quaity indicators that matter to them. This is the cornerstone of a patient-centred heath care system, as opposed to a soey technicay-centred system. v Patients Experiences: Top Heavy with Research 1

6 1 Introduction There is a substantia body of work exporing heath care services from patients perspectives. Athough traditiona measures continue to pay an important roe in evauating heath care quaity, greater emphasis is now being paced on patient-reported outcome measures such as: Patients satisfaction Patients experiences Patients perceptions Patients attitudes This iterature review is not a systematic review of the iterature on patients experiences but rather a critica review of some recent studies the methods used, assumptions made and concusions drawn. The aim of this iterature review is to: Carify types of feedback and different constructs measured Investigate effective ways for coecting feedback Determine variabes that determine patients experiences Appraise the different ways patients feedback is used The objective of the iterature review is to inform the design of a study to deveop a more nuanced understanding of peope s experiences of primary heath care services in the Bayside Medicare Loca area. 1.1 Literature search A semi-structured search of academic databases was undertaken. The databases used were Medine and PsychInfo. The search was imited to artices pubished The search used a series of recognised terms such as primary heath care, patient and consumer satisfaction and experience. The search terms were grouped to form a strategy designed for maxima retrieva of reevant studies in the database. A searches were imited to artices in Engish. The search strategy identified a tota of 2106 abstracts: 1908 from MEDLINE and 208 from PsychInfo. The abstracts were read by a singe reviewer and coded thematicay. A seection of artices from each theme was criticay reviewed. In addition to the automated search strategy, Googe and Googe Schoar were used to search for unpubished reports ( grey iterature ) and patient-centered experience websites. Finay, the reference ists of the incuded studies were checked to identify additiona eigibe references. In tota, 114 artices were incuded in the iterature review. These artices incude research studies, systematic reviews, expert opinion and government reports. v 2 Patients Experiences: Top Heavy with Research

7 Patient feedback A key component of patients wiingness to provide feedback is a desire to improve things for other patients (Brown et a. 2009). Patients feedback is mosty coected through surveys and questionnaires. However, exacty what is being measured in these surveys, or why it is being measured, is not aways cear (Edwards et a. 2011). 2.1 Terminoogy Cornwe and Goodrich (2011) caim that this area of research is bedevied by mutipe terms with over-apping but different meanings (p1). They caim the efforts of researchers, bureaucrats and practitioners to understand and improve patients experiences have been undermined by mudded thinking (p1). The most obvious exampe of this mudded thinking is the incorrect way in which the terms patients satisfaction and patients experiences are used. These two terms are often used interchangeaby when in fact patients satisfaction and experiences are two distinct constructs. In the UK, the definition of patient experience varies from trust to trust, and study to study (Robert et a. 2011). The Inteigent Board (2010) suggests that these constructs are confused because there is not a universa definition of patients experience or patients satisfaction. Robert et a. (2011) caim that it is important to understand the definition of these two constructs Definition of patients experience The Inteigent Board (2010) offers the foowing definition of patient experience: Patient experience is feedback from patients on what actuay happened in the course of receiving care or treatment, both the objective facts and their subjective views of it. The factua eement is usefu in comparing what peope say they experienced against what an agreed care pathway or quaity standard says shoud happen. The opinion eement tes you how patients fet about their experience and heps to corroborate (or otherwise) other quaity measures (p7) Definition of patients satisfaction Satisfaction is a compex, mutidimensiona construct though it is often measured as if it were unidimensiona (Maus et a. 2011; Kaucy et a. 2009). Some caim satisfaction indicates whether or not patients expectations were met (The Inteigent Board 2010); others caim expectation may not be a predictor of satisfaction (Stenberg et a. 2012). A quaitative study found patients do not share a definition of satisfaction (Marcinowicz et a. 2010). Interestingy, Hush et a. (2010) found patients coud be satisfied with their heath care without any improvement in their heath status. This suggests that a patient s cinica outcome is not aways a determinant of patients satisfaction. 2.2 Types of patient feedback There are many different types of patient feedback. This iterature review focuses on two types of feedback: (1) patients satisfaction surveys and (2) patients experiences. Studies that track patients visits with heath care practitioners are aso reviewed Patients satisfaction surveys Evidence suggests that most peope are satisfied with their heath care regardess of the quaity of the care they receive even those who have negative experiences are satisfied with the care they received (Worth 2013; Haggerty 2010; Kaucy et a. 2009). This is particuary the case for oder peope. A US study of oder patients found that oder patients eve of satisfaction with the quaity of their primary care was not a good measure of the quaity or effectiveness of the primary care service (Mod et a. 2012). In the 1990s, it became evident that patient satisfaction data were probematic (Kaucy et a. 2009). A review of 195 studies on patient satisfaction found that the instruments/ toos used to measure satisfaction acked reiabiity and vaidity (Sitza 1999). In addition, dissatisfaction is non- Patients Experiences: Top Heavy with Research 3

8 Patient feedback specific, making survey findings useess for improving patients experiences (Kaucy et a. 2009; Reeves and Seccombe 2008; Couter 2006). Kaucy et a. (2009) ist the probems with patients satisfaction surveys as: The ack of a universa definition of the term satisfaction A disincination for patients to be critica because of not wanting to jeopardise their treatment Satisfaction being determined argey by factors other than the actua heath care an individua receives Findings from satisfaction surveys being nonspecific. Despite these probems, measures of satisfaction continue to be coected Patients experience surveys Patients experience surveys have begun to repace patient satisfaction surveys to measure the quaity of heath care services. Saisbury, Waace, Montgomery (2010) caim that patients experiences provide a more discriminating measure of a heath service s quaity and performance than questions about satisfaction. Kaucy et a. (2009) ist some reasons for studying patients experiences. These reasons incude: Externa accountabiity of heath care providers Enhancing patient choice Improving the quaity of care Measuring the performance of the heath care system as a whoe. Patients experiences surveys have been adapted from consumer surveys used in marketing (Edwards et a. 2011). They measure aspects of care that are important to patients. However, which aspects shoud be measured and in what ways is a recurring probem (Siwa and Okane 2011). According to Roand (2012), it is important to determine what shoud be measured, how it shoud be measured and what difference the measurement might make (Roand 2012). Brown et a. (2009) agree, arguing that feedback shoud be used systematicay and according to a ceary defined strategy. Surveys measure two distinct aspects of care that determine patients experiences. 1. Functiona aspects of care (e.g. waiting times, access, ceaniness) 2. Reationa aspects of care (e.g. dignity, compassion, emotiona support) Robert et a. (2011) argue that most surveys focus on functiona aspects. They argue that more attention shoud be paid to reationa aspects of patients experiences. In the past, surveys focused on a specific heath care service, not an individua practitioner. As such, they focused on functiona aspects of heath care services, not aspects of the practitioner-patient interpersona reationship. Increasingy, studies focus on patients experiences with a specific practitioner (Burford et a. 2012; Moore et a. 2011; Hueston and Carek 2011; Moore et a. 2011). Kenten et a. s (2010) findings highight that reativey minor aspects of a medica consutation can have a significant impact on the patients experiences. They found that the simpe action of doctors greeting patients and introducing themseves coud make the patient fee more comfortabe. Other behaviours that have aso been shown to affect patients experiences incude how doctors address them (Moore et a. 2011), the cothes that doctors wear (Hueston and Carek 2011) and how patients receive test resuts (Eder and Barney 2012). Athough these behaviours may affect patients experiences, it is unikey these behaviours wi affect patients cinica outcomes Tracking patients visits Patients sef-reports of their number of contacts with heath care services are not aways accurate. In their study of patients who had experienced a stroke, Chishti et a. (2013) found patients under-reported the number of consutations they had with a GP. They argue that researchers shoud consider vaidating a sampe against eectronic records, particuary if patient sef-reports of heath care usage are to be used in economic evauations in primary care. Chishti et a. (2013) acknowedge that obtaining patient records requires more effort than obtaining information from patient questionnaires. Jackson et a. (2012) foowed patients with chronic obstructive airways disease for a three-month period foowing hospita discharge. To assist with accuracy of patients reports of their contact with heath care services, Jackson et a. (2012) asked participants to compete a og. This og kept track of patients contacts with heath care services. The ogs contained detais of each visit with a heath care practitioner such as the reason for the visit. This information may otherwise have been forgotten during an interview. 4 Patients Experiences: Top Heavy with Research

9 Patient feedback Key messages from Section 2 1. Without a universa definition of satisfaction, measurements of patients satisfaction are probematic. 2. Patients experiences provide a more discriminating measure of a heath service s quaity than questions about satisfaction. 3. Reativey minor aspects of a heath care consutation may have a significant impact on patients experiences though not on their cinica outcomes. 4. Patients sef-reports of their number of contacts with heath care services are not aways accurate. v Patients Experiences: Top Heavy with Research 5

10 3 Coecting patients feedback There are many different ways heath care organisations coect data on patients experiences. According to the iterature, the most effective ways to coect meaningfu data on patients experiences are: Mapping customer journeys and coordinating data coection across pathways of care. Coecting feedback from patients on core domains frequenty Deveoping systems and processes to support coection using rea time data Buiding fexibiity for oca organisations/ services/teams to capture ocay reevant issues into data coection toos. (Robert et a. 2011; Foot and Cornwe 2010; Brown et a. 2009) The foowing section describes some technica issues in coecting data, incuding: 1. Samping 2. Timing of data coection 3. Methods 3.1 Samping It is important to design an incusive strategy that wi ensure an adequate sampe size and one that is representative of a those who use the heath care service. Evidence indicates that response rates vary among different groups. Those who fai to respond to surveys tend to be young 1, poor, and uneducated (Kaucy et a. 2009). Other groups aso under-represented incude CALD communities and peope with disabiities. To avoid samping bias, it is necessary to design coection methods that encourage these under-represented groups to provide feedback. It is important to ensure patient feedback is coected on an ongoing basis from a representative sampe of patients, incuding disadvantaged groups (Brown et a. 1 Peope under 18 are often not incuded in sampe. 2009). It is aso important that a sampe incudes patients with sufficient experience with a heath care service to be abe to comment. 3.2 Timing of data coection The timing of data coection is critica to ensure organisations use feedback effectivey (Brown et a. 2009). Feedback coected at different times wi potentiay provide different responses Annua In Austraia and overseas, measurements of patients experiences rey on nationa surveys. These nationa surveys often occur annuay (e.g. ABS Patient Experience Survey; Victorian Patient Satisfaction Monitor) or biannuay (The Menzies Nous Austraian Heath Survey). Data are avaiabe infrequenty Rea-time Rea time feedback refers to coecting data from patients soon after they used a heath care services. Increasingy, organisations are designing rea-time patient feedback systems. Brown et a. (2009) caim that rea-time patient feedback provides organisations with an opportunity to increase their responsiveness to service users. According to Brown et a. (2009), the fresher the information, the more effective it can be. Foot and Cornwe (2010) argue that managers and cinica teams shoud monitor quaity of care as often as they monitor budgets. They need reevant, accurate, timey, frequent information from patients to make improvements and compare their own services with others. This requires access to rea time or near rea time feedback, based on standard questions, with demographic information to aow for assessment of popuation mix, and subsequent case mix adjustment (Foot and Cornwe 2010). Rea time data may cutivate a greater sense of staff ownership; a greater a sense of diaogue between staff and 6 Patients Experiences: Top Heavy with Research

11 Coecting patients feedback patients; and foster a greater interest in the consistency of the quaity of services across an organisation (Robert et a. 2011). However, Robert et a. (2011) argue that rea time data typicay focus on snapshots of individua experiences of care and do not refect the wider context around such episodes. There are a range of products, such as hand-hed devices and touch screen kiosks that coect rea-time feedback. Patients feedback using these products has imitations, not east because they can ony gather responses to what questions. The why and how questions require face-toface methods to better understand the experience of the person. In addition, rea time data coection often does not consider samping and can be unrepresentative (e.g. ikey to excude oder peope). 3.3 Methods A number of different methods have been used to measure patients experiences, dividing broady into quaitative and quantitative methodoogies. Both methodoogies are usefu but for different purposes. Surveys may provide information about what is important to patients (Maus et a. 2011) but they do not indicate why it is important. Surveys do not provide nuanced understandings of peope s experiences of heath care services. Increasingy, mixed methods are being used, such as using both posta, phone and onine surveys. The key to effective data coection is to use mutipe methods and data sources, to enhance representation and therefore the vaidity of research findings Quantitative methods This section discusses a range of different quantitative methods incuding: Sef-administered posta surveys Teephone surveys Onine ratings and surveys Rea time surveys (e.g. hand hed devices, touch screen kiosks) Sef-administered posta surveys Posta surveys are convenient and generay user-friendy, depending on the survey s ength and compexity. Posta surveys are favoured by certain groups of the popuation, such as oder peope and those with ower educationa standards. Anonymous sef-administered posta surveys can resut in higher reporting of a practitioner s undesirabe behaviour that face-to-face methods. It is possibe to obtain arge voumes of quantitative data though the use of posta surveys. The average response rate of posta surveys (with mutipe reminders) is 38%. This response rate is beow the minimum that is recommended for epidemioogica studies (Hush et a. 2010) Teephone surveys Both interviewer-assisted and computer-assisted teephone surveys are used to administer surveys that coect patients feedback. Computer-assisted teephone interviewing (CATI) is a technique in which the interviewer foows a script provided by a software appication. This method is used for arge research studies. The ABS Patient Experience Survey and the Austraian Heath Survey, for exampe, both used CATI. Teephone surveys with CATI are usuay shorter that traditiona teephone interviews. Unike interviewer-assisted teephone interviews, CATI do not give participants an opportunity to provide detaied responses, nor often enough time to provide we considered answers Onine ratings Increasingy, Internet sites provide opportunities for patients to rate their practitioners. Greaves et a. (2012) caim that patients now rate their famiy doctors on the Internet in the same way as they might rate a hote on TripAdvisor or a seer on ebay. There are aso web-based assessment toos such as Goba Rating Scae (GRS) that makes a series of statements requiring users to answer yes or no. From the answers, a heath care service s GRS score is automaticay cacuated. The GRS is used by heath services to assess how we they provide patientcentred services (Sint Nicoaas et a. 2012). Both Greaves et a. (2012) and Lopez (2012) caim that findings from Internet and traditiona paper based survey measures of patients experiences are simiar. However, Greaves et a. (2012) note that unsoicited web based ratings (and comments) are often anonymous, making it impossibe for case mix adjustments. Rozenbum and Bates (2013) argue that peope using website ratings may be more extreme (positive or negative) in their views, and be younger than the genera popuation. Aso, they argue gaming may occur in which providers or their representatives give favourabe ratings to boost the ratings of the heath care service Onine surveys Onine surveys are increasingy being used to assess the quaity of heath care from patients perspective. Zuidgeest et a. (2011) argue that the potentia benefits of onine surveys (e.g. reduced effort, quick and ower costs) shoud be baanced against potentia weaknesses (e.g. ow response rates and ack of accessibiity for those without the Internet). Peope who use the Internet are Patients Experiences: Top Heavy with Research 7

12 Coecting patients feedback more affuent, better educated, more often mae, and younger than peope who do not use Internet (Zuidgeest et a. 2011). Sef-administered onine surveys have the same advantage as paper surveys in that they avoid the moderating effects of an interviewer s presence, and may therefore resut in more truthfu responses to sensitive questions than persona interviews. In their study, Martino et a. (2012) found that a one-page etter, signed by the chief medica officer, emphasising the importance of the onine survey, and a brief phone ca reminder improved the response rate to an onine survey. In popuations that aready use the Internet, onine surveys have been found to be a usefu means of conducting research. They have shorter response times than posta surveys and some onine surveys have shown much higher response rates than posta surveys (Zuidgeest et a. 2011) Rea time surveys Surveys on handhed devices Persona hand hed devices are being used to coect rea time data from patients about their experiences, mosty in hospita. Eastern Heath, for exampe, has impemented a new device that provides instant feedback from patients during their hospita stay (Hendry and Gatehouse 2013; Gatehouse 2011). The Patient Experience Tracker System (PETS) records responses from patients about their experience, incuding treatment from staff, invovement in their own care and communication. PETS is part of Eastern Heath s In the Patient s Shoes strategy, which promotes the need for staff to use patient feedback to inform and improve the way they provide care. Robert et a. (2011) questioned the accessibiity and utiity of the hand-hed devices. Oder patients and those for whom Engish is not their first anguage found these devices difficut to use. They aso expressed some concerns that these devices were given mosty to friendier and more cooperative patients. Some heath care services had tested hand-hed devices, abandoned them and moved on to other methods. Others continued to find them usefu (Robert et a. 2011). Touch-screen kiosks Dirocco and Day (2011) examined the feasibiity of coecting feedback from patients at the point of care using touch screen kiosk technoogy. They argue that kiosks are an important advance, however participants sef-seect which causes a sampe bias Mixed quantitative methods Combining an onine survey with a traditiona paper foow-up survey is being used as an aternative to a posta survey (Zuidgeest et a. 2011). Patients then have a choice of competing the survey onine or as a paper survey (e.g. Victorian Patient Satisfaction Monitor 2012). To encourage patients to compete the Victorian Patient Satisfaction Monitor, patients who had not returned the paper survey or competed the survey onine within two weeks were forwarded a reminder etter. The most recent Victorian Patient Satisfaction Monitor had a response rate of 38%, with 8% competing the survey onine (2012). Zuidgeest et a. (2011) argue that a mixed-mode survey (both paper and onine survey) shoud be used rather than just an Internet survey or just a posta survey. They found that combining an Internet survey with a paper foowup survey was ess expensive than a posta survey. This mixed method aso overcame the probems associated with onine surveys such as the possibe excusion of the edery and ess educated Quaitative methods There are a wide variety of quaitative methods used to expore patients experiences. The most common quaitative methods cited in the iterature are: Compaints and compiments Open-ended questionnaires Face-to-face interviews Teephone interviews Focus groups Web based comments Narrative methods Patient journeys Patient co-design Letters, compiments and compaints Patients generay write etters, comments and feedback cards when they receive exceptionay good or bad care. Compaints and compiments represent both ends of the spectrum. They are not a proactive method of earning systematicay about patients experiences to improve service deivery they rey on patients making the effort to make the compaint/compiment (Kaucy et a. 2009). Nonetheess, compaints and compiments can pay an important roe in evauating the quaity in a heath care setting. Robert et a. (2011) argue that more meaningfu anaysis of compaints is needed. They argue that compaints are typicay not coated or anaysed at oca or nationa eves in a way that is usefu for quaity improvement. 8 Patients Experiences: Top Heavy with Research

13 Coecting patients feedback Open-ended questionnaires Couter (2006) argues that we designed questionnaires for patients coud contribute usefuy to an assessment of both the technica competence and interpersona skis of practitioners. Rather than asking patients to rate their care using genera evauation categories (e.g. exceent, very good, good, fair, poor), Couter suggests asking them to report in detai on their experiences of cinica care during a particuar consutation. These types of open-ended questions are designed to eicit reports on what actuay occurred, and how patients fet about what happened (Couter 2006) Face-to-face interviews Face-to-face methods are perhaps the most incusive method, though do not generate arge numbers. They are effective for gathering feedback and for foowing up on any issues, either cinica or socia. However, face-to-face methods are time consuming and can cost a considerabe amount to administer. Patients are known to ike being foowed up they appreciate having someone check up on them (Cochran et a. 2012). However, foow up interviews can ater the resuts by enhancing patient satisfaction and feeings about continuity of care (Cochran et a. 2012). Face-to-face surveys can be an extremey rich source of data; however the moderating effect an interviewer s presence can have on responses needs to be taken into account. Studies have shown that some peope are reuctant to express concerns openy in face-to-face interviews because they anticipate defensive or hostie reactions from staff or fear an adverse impact on their future care (Entwiste et a. 2003) Teephone interviews There is a distinction between (1) teephone interviews in which respondents are picked randomy; (2) teephone interviews that are pre-arranged with the respondent; or (3) teephone interviews that the respondent expects a foow-up ca at some point after an episode of care. Response rates are ikey to be higher when patients expect to receive a ca. When peope are contacted opportunisticay, teephone interviewing may be viewed as intrusive. Simiar to face-to-face interviews, the presence of an interviewer may produce moderating resuts in peope s responses. They can aso enhance patient satisfaction and feeings about continuity of care Focus groups Focus groups are used as a stand-aone method to expore patients experiences or as a precursor to a questionnaire or survey. Mavaddat (2009) used focus groups as the initia information-gathering phase. These focus groups assisted in the deveopment of a questionnaire that assessed patients views of the quaity of primary care menta heath care services (Mavaddat 2009). Materud and Uriksen (2010), on the other hand, used a focus group to expore obese patients experiences with GPs. They purposefuy seected patients to participate in these focus groups Web based comments Internet sites provide opportunities for patients to document their experiences. Some exampes of websites incude NHS choices, I want Great Care and Patient Opinion. Patient Opinion began in UK in 2005 and is funded by hospitas that subscribe to access the data. The site aows peope to give anonymous onine comments about their experiences of hospita services. Other consumers can then read these comments. The NHS aso coects and shares case studies on the Patient Experience Network website. The aim of the Patient Experience Network is to share ideas to drive improvement in patients experiences. Rozenbum and Bates (2013) describe patient-centred heathcare, socia media 2 and the Internet coming together as the perfect storm. They argue that the Internet and socia media have the potentia to create a major shift in how patients and heathcare organisations connect. Greaves et a. (2013) suggest descriptions of patients experiences on socia networks, bogs, Twitter and hospita review sites shoud be coected and anaysed as a too for continuous service monitoring. They argue that this data coud detect institutiona poor performance immediatey, and in a vaid and consistent way in fact, socia media coud detect poor performance before conventiona measures of heathcare quaity (e.g. annua surveys). They caim that socia media coud capture information on a daiy basis and at ow cost. Greaves et a. (2013) caim that anywhere peope tak about their experience of heath care onine is a potentia source of information (e.g. socia networks, Twitter, Facebook, discussion fora and rating websites). However, to use data from the Internet and socia media is a compex task. It requires harvesting free text and then processing this data into usefu information (Greaves et a. 2013). The process invoves identification of appropriate websites, and then scraping that is, puing reevant information 2 A search on Facebook found pages for Primary Heath Care in both Qatar and Nigeria in which service users made comments about their heath care experiences. Patients Experiences: Top Heavy with Research 9

14 Coecting patients feedback off websites on a reguar, automated basis using speciaised software and then using agorithmic processes such as natura anguage processing (Greaves et a. 2013). Lyes and Sarkar (2013) agree that socia media provides a new way to engage in a diaogue with patients. However, they recommend traditiona quaitative anaysis rather than big data anaytica techniques such as natura anguage processing. They argue that natura anguage processing may miss the nuances in patients experiences (Lye et a. 2013). Rozenbum and Bates (2013) caim that using the Internet and socia media to share heath care experiences is vauabe for those patients who use these patforms, but question the vaue of this data for heath care organisations. Rozenbum and Bates (2013) argue that approaches that use socia media have many potentia biases because they do not come from representative segments of the popuation. They aso warn that socia media can be gamed Narrative methods? Narrative methods invove interviewing patients about a particuar encounter with a heath service. This method aows patients to identify incidents or points in the patient journey that have had a significant impact on the patient. With narrative methods, the researcher can expore persona experiences beyond the boundaries of a questionnaire. Narrative methods have the potentia to provide insights into various heath practices that can hep guide the provision of effective heathcare services (Hsu and McCormack 2010). Narrative methods can eicit detaied information that can be used for quaity improvement for specific aspects of service provision. They give an actua representation of whether best practice standards are met, or organisationa poicies and procedures foowed, without being constrained by a survey. However, the number of patients interviewed is sma, as the one-on-one interview method is a time consuming process for gathering information. Aso, researcher characteristics, incuding respect, understanding, and acceptance, are key attributes of success in conducting narrative work, particuary with oder peope (Hsu and McCormack 2010). According to Robert et a. (2011) patients stories create a sense of ownership and motivate staff to find soutions to probems (Robert et a. 2011). However, a sma sampe size has the potentia for organisations to contest the resuts in the event of poor findings as the findings coud be argued to be unrepresentative. Both Tsianakas et a. (2012) and Petroz et a. (2011) compared surveys and patient narratives. In Petroz et a. s (2011) study, participants noted that the Likert-scae approach required them to aggregate their feedback rather than share their individua stories and perceptions when evauating care. Tsianakas et a. (2012) found survey data hepfu as a screening too to identify potentia probems within the breast cancer service, but did not provide sufficient detai of how to improve the service. Tsianakas et a. (2012) recommend future studies use survey and open-ended questions. Narrative methods are different from traditiona audit processes. Traditiona audits test whether a poicy or procedure exists in an organisation. Narrative methods describe how poicy or procedure is actuay enacted Patients Journeys Jackson et a. (2012) argue that it is important to document patients journeys across program and organisationa boundaries. Patient stories can inform strategies to improve and integrate heath care services. Integration has been identified as having potentia to address many current issues in heath care such as access, continuity of care and quaity (Jackson et a. 2012). Their study, for exampe found patients had concerns about heath care system integration such as: (1) system navigation, (2) access and (3) socia support. Sampe sizes for patients journeys are often extremey sma. Jackson et a. s (2012) study foowed ony four (4) patients with chronic obstructive pumonary disease. These four patients were foowed for three months foowing discharge from hospita. During this period, three (3) interviews were conducted with each participant. In addition, participants were invited to document their contacts with heath care services in a og. The ogs provided information such as date, type of contact (phone or visit), practitioner contacted, reason for contact, contact outcome, and other comments deemed reevant by each participant. Jackson et a. (2012) found these ogs an important addition to the interview process because they enabed participants to te their stories effortessy (p230). Jackson et a. (2012) aso used the ogs to deveop maps that visuay depicted each participant s journey. These visua representations highighted how many appointments and how many different providers each participant encountered as we as the timing of each. The patient journey methodoogy has been considered for accreditation and auditing processes. The Austraian Commission on Safety and Quaity in Heath Care, for exampe, has recenty suggested using patient journey interviews to compement traditiona auditing processes for heath care organisations 3. The Austraian Counci of Heathcare Standards are currenty using face-to-face surveys in the Patient journey surveys Patients Experiences: Top Heavy with Research

15 Coecting patients feedback According to Greenfied et a. (2012), there is itte empirica evidence to support using patient journey methodoogy for accreditation. Their study compared patient journey surveys with the current accreditation survey. They concuded the patient journey surveys compement traditiona methods of accreditation. They suggest further research is required to cacuate the costs and benefits of incuding patient journey surveys within accreditation programs Patient co-design Drawing on concepts from the design sciences, experience-based co-design (EBCD) focuses on how staff and patients move through and interact with different parts of a service. EBCD is a form of participatory action research that seeks to capture and understand how peope experience a process or service. Patients and staff share their respective experiences, identify and agree on improvement priorities and work together to achieve them. In Tsianakas et a. s (2012) study, fiedwork invoved 36 fimed narrative patient interviews, 219 hours of ethnographic observation, 63 staff interviews (receptionists, nurses, doctors) and a faciitated co-design change process invoving patient and staff interviewees over a 12-month period. Key messages from Section 3 1. There are important technica issues invoved in coecting data, incuding the numbers needed to obtain reiabe resuts, the timing of the data coection and the way in which responses vary within different popuation groups. 2. Quaitative and quantitative methodoogies are used for different purposes. 3. No singe method of coecting feedback wi reach every group within the community. 4. An incusive strategy wi invove a number of different methods. v Mixed methods Quaitative data (from compaints, patient stories, focus groups, socia media and observations) can be used with quantitative data on cinica quaity, activity, costs and staff experience. This data can be presented in a way that tes a story: about whether and how the quaity of experience is changing over time; whether it is reiabe across the organisation; and how it compares with the quaity of services in other organisations (Robert et a. 2011). 3 Interestingy, the RACGP has endorsed The Practice Accreditation and Improvement Survey (PAIS) to obtain patient satisfaction data (not patient experience) for accreditation purposes. PAIS is an exit-survey that is administered by the receptionist to a number of consecutive patients. The questionnaire contains 27 items that are rated by patients. It takes about 3-5 minutes to compete. Resuts from patient responses are aggregated and presented in graphic format to the practice according to various patient characteristics such as gender and number of previous visits. The survey has been designed by CFEP, and the iterature to support the survey (cited on the webpage) was pubished 12 years ago. 4 patient-cinician-communication/ Patients Experiences: Top Heavy with Research 11

16 4 Survey instruments In recent years, there has been a proiferation of different instruments/toos that describe, measure or compare patients experiences. These toos use different approaches, formats and questions to measure simiar attributes. Robert et a. s (2011) report did not incude a ist of the different toos that have been used because the ist was too ong the ist extended to 98 pages. Athough many toos ask simiar questions, the wording of the questions are different (Robert et a. 2011). Systematic reviews highight the arge number of different toos used to measure patients satisfaction and experiences. For exampe, Ader et a. s (2010) systematic review incuded 12 studies with 9 different satisfaction measures. The arge number of different toos makes it difficut to compare findings. This expains, in part, why so few studies are eigibe for incusion in systematic reviews. In Hush et a. s (2011) systematic review of patients satisfaction with muscuar skeeta therapy, their search ocated 3,790 citations. However, ony 15 studies met the incusion criteria (0.4%) Simiary, Hudon et a. s (2011) systematic review of patients perceptions of patientcentred care ocated 3,045 artices with 26 meeting the incusion criteria (0.8%) and Ridd et a. s (2009) systematic review of doctor-patient reationship found 1985 abstracts of which 11 studies were incuded (0.5%). Athough there is a pethora of data, it is often difficut to make sense of this data because different surveys ask different questions. Without standardised surveys, it is impossibe to compare findings with other heath services, or often even within the same service over time. Picker Institute Europe pioneered measuring patients experiences of heath care. Picker produces a series of survey toos that have become the foundation of many surveys wordwide. The Picker toos measure eight dimensions of care that are important to patients: Information and education Coordination of care Physica comfort Emotiona support Respect for patients preferences Invovement of famiy and friends Continuity and transition Overa impression of care In USA, the Consumer Assessment of Heathcare Providers and Systems (CAHPS) provides a nationay standardised, vaidated too to measure patients experiences in primary heath care practices. The CAHPS surveys have been vaidated and are readiy avaiabe at no charge in the pubic domain. Additiona question can be added to the core survey (Browne et a. 2010). Patients are asked to assess their experiences in areas that research has shown patients vaue such as: Ease of scheduing appointments Avaiabiity of information Communication with cinicians Responsiveness of cinic staff Coordination between heath care providers. According to Kaucy et a. (2009), surveys of patients experiences in Austraia shoud cover the dimensions of care incuded in the Austraian Charter of Heathcare Rights. These seven rights are: Access Safety Respect Communication Participation Privacy Abiity to comment. In addition, Kaucy et a. (2009) argue surveys of patients experiences shoud cover coordinated/integrated care provision, managing transition and accountabiity. Brown et a. (2009) caim that organisations choose to use different questions in different settings and for different purposes. They aso caim that the questions 12 Patients Experiences: Top Heavy with Research

17 Survey instruments asked in surveys are often the wrong ones. They argue this is due to survey design which is determined by managers and/or researchers, rather than by patients. With the wrong questions, these surveys do not coect the data that is required (Brown et a. 2009). The Doctors Interpersona Skis Questionnaire (DISQ) 5 is frequenty used to assess the practitionerpatient reationship. It has been used in a variety of professiona contexts hospitas, genera practices and community heath (Burford et a. 2012). Surveys ike DISQ focus on communication, a key eement of the cinica consutation (Burford et a. 2011). DISQ contains 12 items with each item answered on a five point scae - Poor, Fair, Good, Very good and Exceent. The items incuded are: Satisfaction with visit Warmth of greeting Listening skis Expanation skis Reassurance Confidence in abiity Abe to express concerns and fears Time in consutation Respect shown Patient s persona context Patient as a person Recommend doctor to a friend. Key Messages from Section 4 1. A arge number of different instruments have been used to describe, measure and compare patients experiences. 2. Without standardised surveys, it impossibe to compare findings from different studies. v 5 The DISQ is owned and operated by CFEP Surveys. It is currenty being used by RACGP. Patients Experiences: Top Heavy with Research 13

18 5 Key variabes that determine patients experiences A compex mix of organisationa and human factors affect patients experiences. For exampe, Burford et a. (2011) suggest different patients may vaue different behaviours and quaities in practitioners. They aso argue that communication between practitioners and patients varies with cinica contexts, and may be infuenced by age and gender differences between patient and practitioner. There is evidence that patients experiences are infuenced by their age, gender, ethnicity, educationa eves, heath status, expectations, disposition, socia status, time since care, and previous experience (Robert et a. 2011; Haggerty 2010; The Inteigent Board 2010; Kaucy et a. 2009). However, the reasons for this are not we understood. The finding that patients characteristics infuence their experiences raises an interesting question. Does this refect different expectations among different types of patients? Or do different types of patients within the same heath service receive different types of care? This question is particuary reevant when payments are inked to surveys findings, as in the UK. Heath services ocated in ow socioeconomic areas, for exampe, coud be disadvantaged by oss of income (Saisbury, Waace, Montgomery 2010). Damman et a. (2011) argue that it is not possibe to make fair comparisons of heath services that use data from patients experiences surveys without adjustments for case-mix (Damman et a. 2011). It is important to note that anonymous data, such as when peope describe their experiences on websites, cannot be adjusted for case-mix. 5.1 Patient characteristics Age Numerous studies confirm that oder patients provide more favourabe perceptions of care than younger patients (e.g. Lyratzopouos et a. 2012; Addink et a. 2011; Kontopanteis et a. 2010; Mead and Roand 2009). Addink et a. (2011), for exampe, found young peope reported the owest eves of satisfaction and experience of access to GPs Gender Stenberg et a. (2012) concuded that gender affected both expectations and experiences, however their anaysis however did not invove gender per say. Their assumption that confident type is more often mae and ambiguous type more often femae is arguabe. Hush et a. (2010) found femae patients reported sighty higher satisfaction than mae patients. They suggest that this difference may be due to different determinants of satisfaction for exampe, femae patients focus on communication whie mae patients focus on treatment outcomes (Hush et a. 2010). The important finding from studies that anaysed gender is that the differences in patients experiences associated with gender are sma (Lyratzopouos et a. 2012; Stenberg et a. 2012; Hush et a. 2010) Ethnicity Lyratzopouos et a. (2012) found Asian patients reported a ess positive primary care experience than white patients. The arge number of Asians who ive in ow socioeconomic areas and attend services that receive ow scores on patients experiences surveys can, in part, expain this finding. However, Lyratzopouos et a. (2012) aso found that Asians reported ess positive experiences of doctor-patient communication than white patients in the same practices Education Rademakers et a. (2012) found that patients educationa eves impacted on their experiences of patient-centred care, though the impact was ony sma. Less educated patients regarded patient-centred care as ess important than educated patients. Less educated patients reported 14 Patients Experiences: Top Heavy with Research

19 Key variabes that determine patients experiences receiving too much, and more highy educated patients too itte, in the domains of communication, information and shared decision making (Rademakers et a. 2012). It has aso been observed that some patients prefer to be more invoved in consutations than others (Burford et a. 2011) Heath status Lyratzopouos et a. (2012) found those in poor heath reported a ess positive primary care experience than those in better heath. This finding is consistent with an earier systematic review (Crow et a. 2002). This difference may be accounted for by those in poor heath having a higher exposure to heath care services, increasing chances of having a negative experience (Lyratzopouos et a. 2012). Key messages from Section 5 1. There is evidence that patients experiences are infuenced by sociodemographic factors (e.g. age, education, ethnicity) 2. To make comparisons of heath services, data from patients experiences surveys need to be adjusted for case-mix. 3. Organisationa factors may aso affect patients experiences. v Work status Addink et a. (2011) found that peope working fu time, or those with ong commuting times to work, reported the owest eves of satisfaction, particuary with access to GPs. 5.2 Organisationa characteristics There are aso organisationa characteristics that affect patients experiences. Boyd et a. (2013) found that patients experiences were affected by the size of heath services, with arger heath services scoring ess than smaer ones. They found that patients experiences of access are better in smaer heath services. Boyd et a. (2013) found that patients in smaer practices were much more positive about their experiences of being abe to get through on the phone, make appointments and see their preferred doctor. Other studies aso show smaer practices are associated with higher satisfaction of continuity of care (Kontopanteis et a. 2010). In addition to size of heath service, research has found a reationship between staff webeing and patient experiences, though this reationship is compex (Maben et a. 2012; Szecsenyi 2011). Szecsenyi (2011) found a correation between doctor s job satisfaction and patients satisfaction. Not surprisingy, peope who enjoy their jobs provide a better quaity of care. Kaucy et a. (2009) aso notes the importance of staff morae and professiona attitudes as factors that affect patients experiences. Other organisationa factors that impact on patients experiences incude division of abour, carity over job boundaries, poicies and procedures and stabiity of staffing (Kaucy et a. 2009). Patients Experiences: Top Heavy with Research 15

20 6 Purpose of patients feedback Draper and Hi (2005) argue that articuating the purpose of patients feedback is a critica first step. However, Edwards et a. (2011) suggest that the purpose of patients experiences surveys remains uncear. It is not cear whether surveys are intended to evauate the individua practitioner, organisation or the entire heath care system (Edwards et a. 2011). Kaucy et a. (2009) suggest that information about patients experiences can be used in many different ways for exampe auditing an organisation against its poicies and procedures, auditing an organisation against externa accreditation or best practice standards, or identifying incidents or points in the patient journey that have a significant impact on the patient. They recognise that patients experiences often invove more than one heath care service (Kaucy et a. 2009). The next section describes the different reasons for coecting patient feedback that are cited in the iterature. These reasons are: 1. Provide information 2. Measure the quaity of a heath service 3. Improve quaity of heath care 4. Change professiona practice 5. Evauate innovations 6. Improve quaity of specific interventions 7. Compare heath services 8. Measure coordination of care 9. Rate GP practices 10. To improve patients compiance 6.1 Provide information Kaucy et a. (2009) caim information about patients experiences can highight aspects of a particuar heath care service that are important to those who use it. They aso caim that patients experiences can inform governments about how adequatey the heath care system is meeting the needs of the popuation. This raises the question about how heath care services, and indeed governments, respond when patients experiences data informs them that the heath care service (or system) is not meeting the needs of its popuation. Brown et a. (2009) argue that the exercise of coecting feedback, and discussing the findings at meetings is ony as good as the action that comes out of it. However, when tabing findings of patients experiences surveys at meetings, the minuted action is often to make a record of the report but to take no further action (Robert et a. 2011). The numerous reports on patients experiences provide information, but not action. Robert et a. (2011) caim that exampes in which patient experience data is used to spark debate and action in meetings are rare. 6.2 Measure quaity of heath care services What constitutes quaity of a heath care service, and how to measure it, is the subject of ongoing debate (Gardner and Mazza 2012). Athough there are varying definitions of quaity in heath care, most definitions now incude patients experiences. According to Kaucy et a. (2009 p1), without systematic ways of coecting information about patients experiences of the heath care system, a vita perspective is missing from efforts to improve the quaity of care. It is arguabe whether patients experiences are a reiabe indicator of quaity. Saisbury, Waace and Montgomery (2010), for exampe, suggest that it is difficut to know whether patients experiences refect differences between quaity of practices, performance of practitioners, or variations between patients themseves. In addition, Rao et a. (2006) concuded that patients experiences are not a usefu measure of the quaity of care, as patients experiences are ony weaky reated to technica and cinica indicators of good care. Despite these reservations, patients satisfaction and experiences are increasingy being used as an indicator of the quaity of a heath care service. Other indicators 16 Patients Experiences: Top Heavy with Research