HARRISON R., WALTON M., MANIAS, E., SMITH–MERRY, J., KELLY, P., IEDEMA, R., & ROBINSON, L. (in press) Using patients’ experiences of adverse events to improve health service delivery and practice: Protocol of a data linkage study of

HARRISON R., WALTON M., MANIAS, E., SMITH–MERRY, J., KELLY, P., IEDEMA, R., & ROBINSON, L. (in press) Using patients’ experiences of adverse events to improve health service delivery and practice: Protocol of a data linkage study of Australian

Please download to get full document.

View again

of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.

Art & Photos

Publish on:

Views: 7 | Pages: 8

Extension: PDF | Download: 0

  Using patients ’  experiences of adverseevents to improve health servicedelivery and practice: protocol of a datalinkage study of Australian adults age45 and above Merrilyn Walton, 1 Jennifer Smith-Merry, 2 Reema Harrison, 1 Elizabeth Manias, 3 Rick Iedema, 4 Patrick Kelly 1 To cite:  Walton M, Smith-Merry J, Harrison R,  et al  .Using patients ’  experiencesof adverse events to improvehealth service delivery andpractice: protocol of a datalinkage study of Australianadults age 45 and above. BMJ Open   2014; 4 :e006599.doi:10.1136/bmjopen-2014-006599 ▸  Prepublication history forthis paper is available online.To view these files pleasevisit the journal online(http://dx.doi.org/10.1136/ bmjopen-2014-006599).Received 11 September 2014Accepted 24 September 2014For numbered affiliations seeend of article. Correspondence to Professor Merrilyn Walton;merrilyn.walton@sydney.edu.au ABSTRACTIntroduction:  Evidence of patients ’  experiences isfundamental to creating effective health policy andservice responses, yet is missing from ourknowledge of adverse events. This protocol describesexplorative research redressing this significant deficit;investigating the experiences of a large cohort ofrecently hospitalised patients aged 45 years andabove in hospitals in New South Wales (NSW),Australia. Methods and analysis:  The 45 and Up Study is acohort of 265 000 adults aged 45 years and abovein NSW. Patients who were hospitalised between 1January and 30 June 2014 will be identified fromthis cohort using data linkage and a random sampleof 20 000 invited to participate. A cross-sectionalsurvey (including qualitative and quantitativecomponents) will capture patients ’  experiences inhospital and specifically of adverse events.Approximately 25% of respondents are likely toreport experiencing an adverse event. Quantitativecomponents will capture the nature and type ofevents as well as common features of patients ’ experiences. Qualitative data provide contextualknowledge of their condition and care and theimpact of the event on individuals. Respondentswho do not report an adverse event will report theirexperience in hospital and be the control group.Statistical and thematic analysis will be used topresent a patient perspective of their experiences inhospital; the characteristics of patients experiencingan adverse event; experiences of informationsharing after an event (open disclosure) and theother avenues of redress pursued. Interviewswith key policymakers and a document analysiswill be used to create a map of the currentpractice. Ethics and dissemination:  Dissemination via aone-day workshop, peer-reviewed publications andconference presentations will enable effective clinicalresponses and service provision and policyresponses to adverse events to be developed. BACKGROUNDAdverse events are a significant problem Preventable harm in hospitals due to health-care activities is one of the top six health pro-blems in the developed world. 1  At least onein nine hospitalised patients suffer anadverse event (AE) which may require extra-care or cause permanent harm or evendeath. 2 It is estimated that AEs add 13 – 16%to hospital costs alone — at least one dollarin every seven spent on healthcare. 3 This Strengths and limitations of this study ▪  The use of data linkage is a novel strategy thatwill facilitate identification of a large number ofrecently hospitalised patients; patient experiencestudies to date consistently suffer from smalland unrepresentative patient samples. ▪  Linkage with admitted patient data allows us totriangulate patient reported experiences withinformation in their medical record for example,the health problem they presented with, howlong they were in hospital and the treatmentsreceived; this information validates the self-reported patient data. ▪  The 45 and Up Study cohort does not include arepresentative sample of culturally and linguistic-ally diverse (CALD) participants. We address thisweakness by analysing a subsample of data ofCALD participants to explore whether theirexperiences differ from the wider sample. ▪  The patient sample only includes those who are45 years or older therefore is not a representativepopulation sample. Adults aged 45 and aboveare more likely to be hospitalised and have morefrequent contact with the health system; there-fore knowledge of the experiences of this groupis valuable. Data from younger patients may beexplored in future work using this study as amodel. Walton M,  et al  .  BMJ Open   2014; 4 :e006599. doi:10.1136/bmjopen-2014-006599  1 Open Access Protocol group.bmj.comon November 25, 2014 - Published by http://bmjopen.bmj.com/ Downloaded from   󿬁 gure does not include human costs such as pain andsuffering or loss of independence and productivity forpatients and their carers, or costs of litigation and settle-ment of medical negligence claims. Measurable and sus-tainable improvements in quality and safety are yet to berealised. 4 Patient symptoms are often indicators of the amount of harm arising from healthcare, 1  with retrospectivemedical record review studies detailing a large rate of errors of commission and omission; many causing  AEs. 5 6  While a very low percentage of patients with AEsinitiate a formal complaint, signi 󿬁 cant numbers (23 – 50%) report concerns related to an AE or undesirableevents. 7 – 11 Complaints from patients may have promptedthe initial scrutiny of safety in healthcare, but patientshave been largely ignored as a source of safety evidenceor measurement and are rarely involved in reporting  AEs outside symptoms. 4 5 12 Little is known about the impact of AEs on patients Health systems rely on patient experiences as a centraland integral source of knowledge of health issues, policy development and service planning for almost all aspectsof healthcare, yet we know little of their experiences, which remain largely absent from any input into betterunderstanding the nature of AEs. Models of the impact of AEs on patients are lacking due to lack of research.Evidence-based models generated from patient reporteddata regarding the types of harm that patients experi-ence is urgently required as these are the patient groupsmost vulnerable to harm.Knowing and understanding patients ’  experiences of  AEs is crucial for creating and maintaining  trust  in theirhealth providers and in the health system. 13 – 16 Patient experience data of AEs will also contribute to under-standing the patient populations such as older patients who may be at  a greater risk of AEs or of negative seque-lae from AEs. 17 – 19 These data will provide insight into AEs that occur after discharge (eg, infection); somepostdischarge AEs are also believed to be due to de 󿬁 -cient handover of information but data are limited. 20 – 22 Finally, patient experience data will contribute to under-standing the association between an AE and patients ’ subsequent health needs (eg , requiring hospitalisation), which has not been explored. 23 AE reporting in Australia  All Australian states and territories have active incident reporting systems; in 2007 the Australian state of New South Wales (NSW) implemented the Incident Information Management System (IIMS). This systemrecords data reports on  ‘ incidents ’  (including AEs along  with other untoward incidents such as accidents orthefts) and after analysis are reported annually. The rela-tionship between error and harm in healthcare iscomplex and the focus in these systems on encouraging staff to report errors or  ‘ incidents ’  is limiting. 24 Most reported errors are errors of commission, althoughchart reviews suggest t hat  acts of omission are implicatedin twice as many AEs. 9 25 There are signi 󿬁 cant levels of under-reporting in current voluntary reporting systems.(ref. 26, p.56) Doctors, for example, rarely report using incident reporting systems resulting in particular eventsdominating such as procedures by nursing staff or nurse witnessed AEs. 27 28 Incident reports have thus been cri-tiqued as  ‘ a non-random sample of identi 󿬁 ed hazardsfrom a larger unknown universe of hazards ’ . 29  Another limitation to ful 󿬁 lling the bene 󿬁 ts of AEsusing incident reporting alone is  ‘ the limited amount and variable quality of the information found withinindividual incident reports ’ . 30  A perception also existsamong health professionals that only AEs with seriousoutcomes should be reported. There is usually no oppor-tunity for patients to contribute to information conveyedthrough incident reporting thereby excluding theirknowledge and experience in models of routine incident reporting. Our research will provide an understanding of AEs through the patient  ’ s own experience of AEs.This focus will add an important dimension to current clinical, health service and policy responses to AEs. Open disclosure Open disclosure is the requirement to provide honest explanations to patients and families who have beenimpacted by AEs. The requirement for open disclosurehas been formalised in Australian healthcare settings,predominantly through endorsement of the NationalOpen Disclosure (OD) Standard in the different statesand territories. 9 There is a paucity of evidence regarding the numberof AEs that are disclosed to patients; one study of mis-takes by junior doctors reported a low 12% disclosurerate of AEs to patients. 31  A US survey found that only one-third of patients who had experienced an AE hadbeen informed; but in another US study eliciting patient-identi 󿬁 ed AEs, 40% reported disclosure (de 󿬁 nedas a positive answer to the question:  “ Did anyone fromthe hospit al explain why the negative effectsoccurred? ” ). 32 33 Evidence suggests that disclosure of  AEs doubled the odds that patient s would give higherratings to the quality of their care. 20 Aim The aim of this research is to investigate their experi-ences of AEs and to identify data that can be used tocreate more effective service and policy responses tosuch events. Objectives 1. To determine the patient experience of AEs, includ-ing patients ’  experience of information sharing (dis-closure) after an AE and the role of patients inreporting AEs.2. To ascertain the frequency and characteristics of AEsexperienced by patients in hospitals located in thestate of NSW, Australia. 2  Walton M,  et al  .  BMJ Open   2014; 4 :e006599. doi:10.1136/bmjopen-2014-006599 Open Access group.bmj.comon November 25, 2014 - Published by http://bmjopen.bmj.com/ Downloaded from   3. To describe the characteristics of patients who experi-ence AEs, in comparison to patients who do not experience AEs.4. To carry out a detailed examination of present service and policy structures designed to deal with AEs and make recommendations for change basedon our  󿬁 ndings.5. To undertake community consultation about the 󿬁 ndings and transfer project   󿬁 ndings to the commu-nity, policymakers, health practitioners and servicemanagers. METHODSEthics approval Ethics approval was granted by the NSW Population andHealth Services Research Ethics Committee (HREC/13/CIPHS/66 &2013/12/496). It was agreed between thetwo ethics committees that the latter ethics committee would be the single committee managing and monitor-ing the research. Design This is a mixed methods study involving data collection via cross-sectional survey from a large research cohort,interviews with high-level policymakers, and data linkagebetween The Centre for Health Record Linkage(CHeReL), the Admitted Patient Data Collection(APDC), the Register of Births, Deaths and Marriages(RBDM) and the 45 and Up Study databases. Setting and participants Study population -survey The sample will be drawn from the 45 and Up Study cohort. The 45 and Up Study is the largest cohort study in the Southern Hemisphere, including 265 000 partici-pants across NSW who are aged 45 and above. The study completed recruitment in 2008 and all recruits haveconsented to the use of their data in research. A widerange of data is available on the 265 000 participants inthe 45 and Up Study cohort, including age, postcode,education, ethnicity, lifestyle and habits, current medica-tions, histor y of disease, surgical procedures and employ-ment status. 34 The choice to use an existing collectionof participants was made because the dif  󿬁 culty of recruiting respondents through their health services is well established. 35  Advertising through the public mediadoes not necessarily attract a representative patient population. Nor do hospital selected patients meet cri-teria for an unbiased study cohort. A primary bene 󿬁 t of the 45 and Up sample is that it willimprove access to a general and guaranteed sample popu-lation. Past efforts to research this topic have been ham-pered by the dif  󿬁 culty of accessing an unbiased samplepopulation. Accessing the appropriate survey populationis an important consideration for this research. Becausethe majority of patients hospitalised in Australia are ov er45 the use of this study population is ideal. 36 Furthermore, AEs tend to occur in older people. In one Australian study (N=979 834), patients with an AE wereolder (M=62.5, 95% CI 62.7 to 63 years) than t hose without an AE (M=48.2, 95% CI 48.1 to 48.3 years). 37 Inanother Australian study (N=1177), age >70 years wasfound to be a strong predictor (OR 1.9, 95% CI 1.3 to2.6) of AEs following surgery. 38 Our study participants will be a random sample of 20 000 from the 45 and UpStudy cohort who were hospitalised at any hospital inNSW over the most recent 6 month period to the time of data collection. These participants will be identi 󿬁 edusing data linkage provided by the Centre for Healthrecord Linkage (CHeReL) with the Admitted Patient Data Collection, which is administered by NSW Health. Limitations of survey sample  A recognised limitation of the 45 and Up Study is that it is not representative with respect to individuals from cul-turally and linguistically diverse (CALD) backgrounds.For example, while only 25% of the 45 and Up Study  were born outside of Australia, 2006 census data put s t his 󿬁 gure at 39% for those aged 45 and above in NSW. 36 Inorder to address this limitation we will analyse a subset of the surveys we receive from CALD participants tocompare the experiences of those CALD participants who experienced an AE with those who did not experi-ence an AE. We will also compare the CALD participantsin the 45 and Up cohort to the non-CALD members of the cohort to see if there are speci 󿬁 c differences or varia-tions in the characteristics associated with the AE. Survey sample size  We anticipate there to be 20 000 participants in thecohort eligible to take part in the study based on thenumber of participants within the cohort who have beenhospitalised. In a 6-month period (July  – December) in2007, there were 18 460 public hospitalisations among 45 and Up Study participants. Hospitalisation rates willrise every year as the cohort ages; thus, we estimate20 000 public hospitalisations (our study population) forthe 6-month period January 2014 –  June 2014. We antici-pate a 60% response rate to our questionnaire based onthe response rates observed in other substudies of 45and Up participants, yielding an estimated 14 000respondents. We expect approximately 3500 respondents(25%) will have experienced an AE based on patient AEself-reporting   󿬁 gures reported elsewhere; the remaining 10 500 respondents are expected not to have experi-enced an AE and will be the control group. Study population — interviews  We will conduct interviews with 30 individuals who havea direct role in managing AEs and thereby holding embodied knowledge about practice in the  󿬁 eld.Sampling will be purposive, based on the roles that indi- viduals ful 󿬁 l. Participants will be initially identi 󿬁 ed by the project reference group associated with the project and subsequently through Clinical Governance Units of  Walton M,  et al  .  BMJ Open   2014; 4 :e006599. doi:10.1136/bmjopen-2014-006599  3 Open Access group.bmj.comon November 25, 2014 - Published by http://bmjopen.bmj.com/ Downloaded from   Local Health Networks. We expect to conduct around30 interviews, including 10 working in policy settings(with clinical governance managers working in govern-ment policy and within key bodies such as the AustralianCommission on Safety and Quality in Health Care, NSW Clinical Excellence Commission and the NSW HealthCare Complaints Commission) and 20 with individuals working in practice settings (service managers, practi-tioners working as complaints managers). We will pur-posively seek to include a sample from two MetropolitanLocal Health Districts (Sydney, Western Sydney) and twoRural and Regional NSW Local Health Districts (Far West, Hunter New England). The  󿬁 nal number of inter- views will be determined as the interviews progress.Interviewing will be halted at the point of data saturationonce no new data are emerging. Procedure The study will run from July 2013 to July 2016 with datacollection starting November 2014. This study comprises 󿬁  ve phases:1. Questionnaire development;2. Completion of contractual arrangements with datalinkage partners;3. Administration of questionnaire;4. Analysis of survey responses;5. Mapping of existing formal service and policy responses to AEs;6. Transfer of  project   󿬁 ndings to the policy and practicecommunity. 39 Phase one — questionnaire development In phase one, the survey for patients was developed andre 󿬁 ned in a series of stages. Past patient experienceresearch was used to develop items to determine whether patients experienced an AE and opportunitiesfor patients to report AEs (during care, discharge, post-hospitalisation). Iedema  et al  11 (chief investigator B) and Weingart   et al  ’ s 40 surveys of patient experiences of opendisclosure were a signi 󿬁 cant  f oundation of data for thedevelopment of these items. 5  We included the Picker Adult Inpatient Questionnaire (PAIQ) which is a vali-dated questionnaire designed to elicit patients ’  ov erallhospital experiences about their care and treatment. 41 It has a high degree of face validit  y, construct validity andinternal reliability consistency. 41 Items about open dis-closure and subsequent decision making about com-plaint or litigation were also incorporated. A validatedinstrument for assessing open disclosure by CI Iedema was used to assess the quality of formal open disclosureagainst new national indicators. Items from this instru-ment along with that used by Weingart in the USA wereused to identify what information was provided about the AE; what action was taken to remedy the situation; what did not happen; intention of making a formal com-plaint or of seeking compensation; areas not addressedthat the patient wanted addressed. We also includeditems to determine where and when the AE(s) occurred(ie, public/private hospitals, in/out NSW, within correct time-frame parameters etc in order to identify possibleconfounding variables). An expert panel reviewedseveral iterations of the survey to attend to issues such aslength, content validity, scaling responses and potentialfor response bias.The resulting survey comprises  󿬁  ve parts A  – E. Part A asks patients for details of their hospitalisation followedby an additional four further validated sections. Part Bincorporates questions about care and treatment fromthe validated Picker questionnaire. 41 Part C includes aseries of items reg arding the healthcare incident based on Weingart   et al. 5 40 Part D asks about experiences of disclosure based on work by Iedema  et al. 11 Part Eexplores patient reports based on items developed by Daniels and also work by Walton  et al   (chief investigator A). 42 The survey was reviewed by the Project ReferenceGroup for suitability and face validity and approved by the 45 and Up Study Management Committee. Phase two — questionnaire administration Phase 2 will include administration of the questionnaireto a sample of recently hospitalised patients from the 45and Up data bank. The Centre for Health RecordLinkag e (CHeReL) is a NSW Ministry of Health Agency. 43 CHeReL will link data from the AdmittedPatient Data Collection with the 45 and Up Study Database to identify participants who have been hospita-lised in the 6-month period between January and June2014 using admission data. A random sample of 20 000 of the eligible 45 and Up participants will be sent a survey pack by the mailing house including an invitation letter, aparticipant information lea 󿬂 et and the survey with a per-forated consent form attached to it. In keeping withother studies of the 45 and Up participants, CALD or any other participants who have dif  󿬁 culty reading in Englishare advised to seek help from a friend or relative tounderstand the study materials and to complete thesurvey. Those who wish to participate are required tocomplete the consent form and return with their com-pleted survey to the 45 and Up Study using the reply-paidenvelope included in the pack. Participants who have not experienced an AE will only complete parts A and B of the survey, providing details about their hospital stay andtheir experience of this stay. Participants who have experi-enced an AE will complete a further three parts (C, Dand E) in which they will answer questions about the AE,the Open Disclosure process and whether they made acomplaint. Returned surveys will be screened by the 45and Up Study team to ensure the consent form is com-plete and then sent on to the research team in ade-identi 󿬁 ed form for analysis. Any surveys with nosigned consent form will not be included. Phase three — analysis of questionnaire data Phase 3 will involve qualitative and quantitative dataanalysis of completed questionnaires to obtain 4  Walton M,  et al  .  BMJ Open   2014; 4 :e006599. doi:10.1136/bmjopen-2014-006599 Open Access group.bmj.comon November 25, 2014 - Published by http://bmjopen.bmj.com/ Downloaded from   comprehensive information on patient characteristics,the nature and frequency of any AEs experienced andthe impact of AEs on patient outcomes, whether thepatient experienced an Open Disclosure process(formal or informal) and whether the patient made acomplaint or initiated legal action. Qualitative analysis  Two analyses of qualitative questionnaire data will beperformed:The conceptual framework of the Internat ionalClassi 󿬁 cation for Patient Safety will be employed. 42 TheInternational Classi 󿬁 cation for Patient Safety categorisespatient safety information using standardised sets of con-cepts to facilitate  ‘ description, comparison, measurement,monitoring, analysis and interpretation of information toimprove patient care, and for epidemiological and healthpolicy planning purposes ’  (ref. 44, p.1). Each AE incident  will be independently classi 󿬁 ed by two safety  experts.Using the method devised by Runciman  et al  45 ‘ naturalcategories ’  will be determined and contributing factors,detection, mitigating factors, patient and organisationaloutcomes and ameliorating factors will be sought. Thecomments that patients make about their AE will beexamined and thematically coded using discourse analysisas this has been effectively used for analysing patient experience in previous research. 46 The analysis will bestructured around the research questions. NVivo will beutilised to manage and track the discourse analyticalprocess that is applied to the data. Patient identi 󿬁 edopportunities for reporting AEs will also be examinedand thematically coded, paying special attention to thetype of AE, location in the hospital, age group, culturalbackground and the opportunities to report. 47 Quantitative analysis  Estimates of the frequency and type of AE identi 󿬁 edfrom questionnaire results will be conducted. The use of the WHO International Classi 󿬁 cation for Patient Safety to categorise patients ’  descriptions of their AEs willensure that foreseeable and known complications of conditions and treatments are separately identi 󿬁 ed anddescribed. Patient experiences that fall outside the classi- 󿬁 cation system will be identi 󿬁 ed and explored.Descriptive statistics (eg, means, SDs, frequencies andpercentages) will be calculated to summarise data col-lected. In particular, estimates of the frequency and typeof AE will be estimated with 95% CIs. Logistic regression will be used to determine if patients who experience AEs differ from patients who do not experience AEs.Potential risk factors will focus on patient characteristicssuch as age; gender; region; chronic disease(s), lengthof stay, transfer to other healthcare services. Otherfactors may be identi 󿬁 ed through the conceptual frame- work noted above. Appropriate model building strategiesand model checking will be employed. 48 Phase 4 — mapping of existing service and policy responses In phase 4 we will analyse existing service and policy responses to AEs by undertaking a mapping exercise,combining data from a detailed document mapping exercise and interviews with key actors in the  󿬁 eld. Theresults of this analysis to make detailed recommenda-tions about how these service and policy structures canbest make use of the patient experience data. Thismapping is an important component of the researchproject because government organisations are mandatedto implement processes related to AEs. An understand-ing of current practice will allow our recommendationsto  󿬁 t into and build on existing practice.Stakeholder interviews: Interviews will be semistruc-tured and open-ended. Questioning will focus onpresent service and policy structures designed to deal with AEs and the extent to which patient experiencesare utilised as a source of knowledge in service andpolicy design. Participants will also be asked which docu-ments they draw on in their work. They will be con-tacted directly via email and asked to participate in aninterview which will take place at their place of work, orby phone if they prefer. Interviews will be recorded withthe participant  ’ s permission and transcribed.Interview Data Analysis: Data will be hand codedthrough the qualitative data analysis software NVivo.Thematic analysis will be conducted based around theproject research aims. Coding reliability will be improvedthrough independent coding of a subsample of theinterviews by different members of the research team.Initial thematic codes from all interviews will be brought together and recoded to draw out subcodes. Individualsections of the data will be freshly coded to ensure analignment between initial and later coding. Singleinstances of codes will be removed and those with closesimilarity to other codes will be merged. A narrative of the themes derived from the coding will be created.Representative quotations will be selected to demon-strate the themes where necessary.Document Analysis: The aim of the documentary ana-lysis will be to  󿬁 nd and summarise the formal docu-ments which structure practice around AEs. Healthdepartment policies, Guidelines distributed by LocalHealth Districts, quality and safety protocols publishedby the NSW Clinical Excellence Commission and the Australian Commission on Quality and Safety will beidenti 󿬁 ed through a range of approaches: (1) literaturereview  — by locating any formal documents cited by aca-demic sources; (2) reference group — through sugges-tions made by the project research group; (3) interview respondents — through locating any documents spokenabout in the interviews. Documents will be summarisedaccording to purpose and audience. Phase 5 — transfer of research findings. Phase 5 will involve transfer of   󿬁 ndings to the commu-nity, policymakers, service managers and agencies. We Walton M,  et al  .  BMJ Open   2014; 4 :e006599. doi:10.1136/bmjopen-2014-006599  5 Open Access group.bmj.comon November 25, 2014 - Published by http://bmjopen.bmj.com/ Downloaded from 
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks