NIHR Greater Manchester PSTRC Plain English Publication Summaries 2014 17 by NIHR Greater Manchester PSTRC

Greater Manchester Primary Care Patient Safety Translational Research Centre

2014-17

Plain English

Publication Summaries

Publication summaries The first NIHR Greater Manchester PSTRC published a range of papers in medical journals from 2014-17. In this booklet, you will find all the plain English summaries that were produced to accompany the papers, listed in date order.

Tools for Primary Care Patient Safety: A Narrative View

4-5

Relationship continuity: when and why do primary care patients think it is safer?

6-7

Trust, temporality and systems: how do patients understand patient safety in primary care? A qualitative study

8-9

Adoption of Clinical Decision Support in Multimorbidity: A Systematic Review

10-11

Application of a tool for the evaluation of public and patient involvement in research

12-13

Contributory factors to patient safety incidents in primary care: protocol for a systematic review

14-15

Primary care medication safety surveillance with integrated electronic health records: insights from the UK

16-17

Developing a Primary Care Patient Measure of Safety (PC PMOS)

18-19

Effectiveness of case management for â&#x20AC;&#x2DC;at riskâ&#x20AC;&#x2122; patients in primary care: A systematic review and meta-analysis

20-21

Missed Diagnostic Opportunities and English general practice: a study to determine their incidence, confounding and contributing factors and potential impact on patients through retrospective review of electronic medical records

22-23

Blame the patient, blame the doctor or blame the system? A meta-synthesis of qualitative studies of patient safety in primary care

24-25

Multimorbidity and Patient Safety Incidents in Primary Care: A Systematic Review and Meta-analysis

26-27

The care.data consensus? A qualitative analysis of opinions expressed on Twitter

28-29

Examining variations in prescribing safety in UK general practice: cross sectional study using the Clinical Practice Research Datalink

30-31

Sensemaking and the co-production of safety: a qualitative study of primary medical care patients

32-33

Non-psychotropic medication and risk of suicide or attempted suicide: a systematic review

34-35

Effectiveness of multidisciplinary team case management: difference-in-differences analysis

36-37

When procedures meet practice in community pharmacies: qualitative insights from pharmacists and pharmacy support staff

38-39

An external validation of models to predict the onset of chronic kidney disease using population-based electronic health records from Salford, UK

40-41

Integrated care: theory to practice

Temporal trends in antidepressant prescribing to children in UK primary care, 2000-2015

43-44

Understanding the implementation and adoption of a technological intervention to improve medication safety in primary care: a realist evaluation

45-46

Antipsychotic Prescribing to Patients Diagnosed with Dementia Without a Diagnosis of Psychosis in the Context of National Guidance and Drug Safety Warnings: Longitudinal Study in UK General Practice

47-48

Never Events in UK General Practice: A Survey of the Views of General Practitioners on Their Frequency and Acceptability as a Safety Improvement Approach

49-50

Publication title: Tools for Primary Care Patient Safety: A Narrative Review

Publication details (Vancouver format) Spencer R, Campbell SM. Tools for Primary Care Patient Safety: A Narrative Review. BMC Family Practice. 2014 Jun 5;15:110. doi: 10.1186/1471-2296-15-110.

What was known before your paper was published? Most people who see a doctor or nurse do so in general practice, which is nearly 340 million visits made annually in England (with 2 percent involving a patient safety incident) and approximately 1 billion community prescriptions annually (with 4 percent involving an error). However, research into patient safety in primary care is a developing field with a small but increasing evidence base, compared to patient safety research in hospital settings. Most of the previous research in primary care has focused on reducing prescribing errors in general practice. To improve safety in primary care settings, we need to know what methods and tools are available to measure patient safety.

What did you do? Our aim was to identify tools described in the scientific literature that can be used by general practitioners as part of a patient safety toolkit to improve the safety of the care and services provided by their practices. We looked at the evidence in 6 healthcare databases in 2011 searching for literature on primary care patient safety. We searched for literature that described three elements: 1. setting (primary care, e.g. general/family practice, ambulatory care, community care, generalist care) 2. safety, including safety synonyms (e.g. error, adverse event, fault, malpractice) 3. types of tool (e.g. indicator, survey, guideline, electronic alerts that a safety issue may be present). Both authors looked at all papers, identifying and discussing key themes and tools.

What did you find? We found 114 tools described in publications, with most literature coming from the USA (41 percent) and the UK (23 percent) within the last 10 years. Most literature looked at tools to prevent medication error (55 percent). Other areas where tools to enhance patient safety were described, were related to processes and procedures within the primary care setting (e.g. a general practice) that create what is known as a â&#x20AC;&#x2DC;safety climateâ&#x20AC;&#x2122; (8 percent) and enable adverse event reporting (8 percent), tools to support patient safety as people move between general practice and hospital care (5 percent), measures to aid correct diagnoses (5 percent), the use of computer based informatics tools (4.5 percent) and the role of the patient in patient safety (3 percent). Tools to address diagnostic error and the way lab test results are handled appeared infrequently (<1 percent of total literature) despite their importance. The remainder of literature (11 percent) highlighted tools related to referrals, Out-Of-Hours (OOH) care, telephone care, organisational issues, mortality and clerical error.

What insights/knowledge did you add? This review identified a range of tools that are available for use in general practice to measure patient safety, which is crucial to improve safety. However, many of the tools have yet to be tested to see if they improve quality and safety of care. The lack of suitable tools on tools to help improve diagnosis and handling lab test results, provide priorities for future research. This study was the first attempt to identify tools that could form part of a Patient Safety Toolkit to improve safety outcomes and patient experiences in general practice, in any healthcare system.

Publication title: Relationship continuity: when and why do patients think it is safer?

Publication details (Vancouver format) Rhodes P, Sanders C, Campbell SM. Relationship continuity: when and why do patients think it is safer? Br J Gen Pract. 2014 Dec; 64(629):e758-64. doi: 10.3399/bjgp14X682825.

What was known before your paper was published? Relationship continuity (seeing the same doctor who knows the patient and the patient knows the doctor) is linked to better patient outcomes and is popular with both patients and GPs. Patients often report higher satisfaction with seeing a doctor they know. However, it is rarely discussed in terms of patient safety, especially from the perspective of patients. Seeing the same doctor has become progressively more difficult with changes in policy and practice organisation (e.g. bigger practices with more doctors and nurses). Research has looked at the patient safety implications of continuity from a professional perspective but patientsâ&#x20AC;&#x2122; views remain largely unexplored.

What did you do? In this project we looked at when and why patients think it is safer to see the same doctor as well as when and why they consider it to be less safe. Most patients were recruited from five general practices in the northwest of England through practice Patient Participation Groups or posters in waiting rooms. Interested patients were sent information about the project and a consent form. We talked to patients using interviews. We then looked at the transcripts of each interview (which could not identify individuals) for key themes and issues.

What did you find? Thirty eight people (14 men, 24 women; age range 18-78 years) were interviewed from 19 rural, small town and city practices. For patients, relationship continuity was not simply a matter of a quality experience but an important safety concern that offered greater psychosocial security than consultations with unfamiliar GPs. Relationship continuity enabled the GP to get to know the patient and foster mutual trust. However, patients were also aware of the risks: a false sense of security and lack of a fresh perspective. Their desire for continuity varied with the nature of their concerns, and perception of GPs' qualities and skills.

What insights/knowledge did you add? Relationship continuity and choice of GP were important safety strategies for patients, neither of which is adequately supported by recent policy changes. Patients want the opportunity to choose relationship continuity when and with whom they feel they need it and are aware of the advantages and disadvantages of seeing the same doctor. Despite its advantages to patients and doctors, relationship continuity appears to be maintained more through patient actions than practitioner actions and has been discouraged by policy initiatives that prioritise speed of access over continuity. No patient supported imposed continuity with a named GP.

Publication title: Trust, temporality and systems: How do patients understand patient safety in primary care?

Publication details (Vancouver format) Rhodes P, Campbell S, Sanders C. Trust, temporality and systems: How do patients understand patient safety in primary care? A qualitative study. Health Expectations 2015 Feb 3. doi: 10.1111/hex.12342

What was known before your paper was published? Patient safety research has tended to focus on hospital settings, although most patient encounters with healthcare providers occur in general practice or in pharmacies. Most research also focuses on errors made by health professionals (e.g. GPs, pharmacists) rather than patients' own understandings and experiences of safety in healthcare. There is little agreement among health professionals about what constitutes error or harm. Studies looking at patientsâ&#x20AC;&#x2122; views reveal broader ideas of both harm and the causes of harm, than those of health professionals. There is a growing acceptance that patient safety is not just the absence of error, not all errors cause harm and not all harms result from error.

What did you do? Our aim was to explore patients' understandings of safety in primary care. We interviewed patients who were recruited from general practices in northern England. Other patients who took part were suggested by participants already recruited. People with more than one health condition were included intentionally as this group is known to be vulnerable to safety incidents. We asked patients for information about themselves and to tell us how they defined patient safety and about their experiences, rather than starting with an existing or professional definition. We looked for common themes and ideas from all those interviewed.

What did you find? Thirty-eight people (14 men, 24 women) from 19 general practices in rural, small town and city locations were interviewed. Patients raised issues about access to consultations with GPs, length of consultations with GPs and being able to see they GP they wanted to see. Three broad themes were found: (i) (ii) (iii)

trust and psychosocial aspects of health professional â&#x20AC;&#x201C; patient relationships (perceptions about whether patients feel their GP trusts them and whether they trust their GP) choice of GP and continuity of relationship with the same GP (the patient being able to see the doctor they want or not, and the advantages and disadvantages of both) organisational issues constraining safety (e.g. lost test results or inaccurate prescriptions).

What insights/knowledge did you add? The study did not uncover a single understanding of what patient safety means to the patients who participated, but a wide range of views that could change with varying circumstances and needs. For patients, safety is not just about access to staff and services but is also but personal issues and this can be influenced by their relationships and interactions with staff. Currently, the way healthcare is delivered does not adequately focus on individual patient needs and experiences and this could be improved. Patients are aware of the risks as well as benefits of different kinds of relationship (e.g. seeing only the same doctor) and seek different types of relationship in different situations.

Publication title: Adoption of Clinical Decision Support in Multimorbidity: A Systematic Review

Publication details (Vancouver format) Fraccaro P, Casteleiro MA, Ainsworth J, Buchan IE. Adoption of Clinical Decision Support in Multimorbidity: A Systematic Review, JMIR Med Inform 2015; 3(1):e4 doi:10.2196/medinform.3503

What was known before your paper was published? Patients with multiple health conditions (known as multimorbidity) have complex needs and the number of people with multimorbidity is increasing as the population ages. This is one of the greatest challenges facing health care. Having more than one condition generates: (1) (2) (3) (4) (5) (6)

interactions between the causes and effects of multiple health conditions duplication of laboratory tests difficulties for doctors and patients in adhering to clinical practice guidelines, which are often conflicting obstacles in the continuity of care confusing patient self-management information errors in the prescribing of medicines.

In this context, clinical decision support (CDS) systems, such as systems that use technology and clinical information to support doctors and patients, need to be able to cope with the complex scenarios that may arise for a patient with multimorbidity and minimise risks of causing harm to those patients.

What did you do? In this project we reviewed the academic literature to investigate to what extent and for which tasks Clinical Decision Support systems were adopted to help with the complex needs that patients with multimorbidity face. After screening thousands of articles by reading the article summary, we selected 50 of them for an in depth fulltext analysis. We finally found 20 articles that matched our inclusion criteria.

What did you find? The prescription of drugs and the integration of different clinical practice guidelines were the predominant clinical tasks that were undertaken within the studies reviewed. The most frequent diseases mentioned were cardiovascular diseases and diabetes mellitus. In all, 12 articles mentioned generalist doctor(s) as the Clinical Decision Support systemâ&#x20AC;&#x2122;s user, showing that multimorbidity is predominantly an issue dealt with in primary care. In the reviewed articles, there were no studies referring to the active involvement of the patient in the decision-making process or to patient self-management. None of the articles reviewed adopted mobile technologies (i.e. smart-phones).

What insights/knowledge did you add? This review of the literature shows that there is a lack of research investigating how technology can be used to support clinical decisions for patients with multimorbidity. Perhaps, the reason for this lack of evidence is the fact that multimorbidity is a relatively new concept in medicine. In fact, there is still a lack of evidence about how different health conditions might interact in different patients. Data to help understand and address the use of informatics to support clinical decision making exists in many electronic health record systems. Utilising this information has the potential to investigate and understand the complexities faced by patients with multimorbidity.

Publication title: Application of a tool for the evaluation of public and patient involvement in research

Publication details (Vancouver format) Stocks SJ, Giles SJ, Cheraghi-sohi S, et al. Application of a tool for the evaluation of public and patient involvement in research. BMJ Open 2015;5: e006390. doi:10.1136/bmjopen-2014-006390

What was known before your paper was published? Involving patients and members of the public in research is crucial. A Research User Group (RUG) was established within the Greater Manchester Primary Care Patient Safety Translational Research Centre (PSTRC), after an open call for recruitment of patients and members of the public to join the group, in April 2013. The aim of the RUG is to fulfil both a strategic governance role and to involve the public and patients in research activities (PPI). Given the high priority of PPI within Greater Manchester PSTRC and research funded by the National Institute for Health Research, it is important that the process and impact of PPI on health services research is evaluated. Although researchers, members of the public, and policy makers believe that it is possible to evaluate the impact of PPI on research, it rarely happens. Our aim was to make some progress towards evaluating the impact of the RUG from the perspective of the RUG members and, in doing so, share a method that could be used to evaluate PPI in other research settings.

What did you do? We used a questionnaire to ask the RUG members to score their personal satisfaction and level of involvement in research approximately every 6 weeks. The way the data was analysed meant that we looked at how the score given by each RUG member changed over time within each person. By within each person, we mean using a mathematical model that takes into account the change in satisfaction and level of involvement scores over time for each RUG member, then combining the measure of change in the scores over time for all the RUG members, to give an overall score for how people have felt over the time that have been part of the RUG. This means that if you were a naturally optimistic person or naturally sceptical it would not affect the change over time and this is a better way to assess how satisfied and involved people feel than just adding up all the scores.

What did you find? Our results showed that the questionnaire did quite well at measuring what was intended in that the questions asking about similar issues had similar responses within each person. It also showed that most of the RUG members gave a high score of satisfaction and level of involvement from the beginning. This was a positive finding but it meant that the scores could only stay the same or go down. This is called a ceiling effect and often happens in this type of situation. In fact there was no overall change in score over the period of the questionnaire but some RUG members gave decreasing scores over time.

What insights/knowledge did you add? We tested a method for measuring how members of the public involved in research feel they are having an effect on the research using a quantitative method (i.e. the opinions of the RUG were scored using numbers rather than words for the first time [as far as we know]). We found that overall the RUG members were satisfied but a small number of RUG members had decreasing scores and it enabled these individuals to be offered extra support. Since publication the method is being tested in other research settings nationally and internationally.

Publication title: Contributory factors to patient safety incidents in primary care: protocol for a systematic review

Publication details (Vancouver format) Giles S.J,

Panagioti M, Hernan A, Cheraghi-Sohi S, Lawton R.L. Contributory factors to patient safety incidents in primary care: protocol for a systematic review Systematic Reviews 2015, 4:63 (7 May 2015)

What was known before your paper was published? In the past, health care organisations have mainly relied on staff incident reporting systems to help them identify instances of when things go wrong. As a result of this a framework, known as the Yorkshire Contributory Factors Framework, was developed to identify the causes/potential causes of error in hospitals. The framework is a questionnaire that is completed by patients to help staff understand whether the care provided in the hospital was carried out safely. This was done by looking at relevant research papers and identifying common themes (a systematic review of the literature) on which the framework questionnaire could be based. We know that eighty-five percent of healthcare contacts occur in primary care (general practice, pharmacy, dentistry etc), and that there is a need for a similar framework that can be used in primary care. This paper describes the process that will be undertaken to develop this framework in a primary care setting.

What did you do? This paper describes the process that will be undertaken for the systematic review. It will involve searching the relevant literature in the field. To do this, we will search health care databases, and the contents pages of some leading patient safety journals. The focus of our search will be any publications that are within a primary care setting, are patient safety focused and contain research findings. Three researchers will be involved in screening and extracting data from the research papers.

What did you find? The results from the systematic review will be published in a separate paper.

What insights/knowledge did you add? This publication has outlined the method that will be used for the systematic review that may form the basis for a new primary care patient safety framework questionnaire. We will be able to share the results of the systematic review once we have completed the systematic review itself.

Publication title: Primary Care Medication Safety Surveillance with Integrated Primary and Secondary Care Electronic Health Records: A Cross-Sectional Study

Publication details (Vancouver format) Akbarov A, Kontopantelis E, Sperrin M, Stocks SJ, Williams R, Rodgers S, Avery A, Buchan I, Ashcroft DM. Primary Care Medication Safety Surveillance with Integrated Primary and Secondary Care Electronic Health Records: A Cross-Sectional Study. Drug Saf. 2015 Jul;38(7):671â&#x20AC;&#x201C;82.

What was known before your paper was published? It is well known that inappropriate prescribing of medicines is a relevant and important issue in healthcare. Up to 59% of medicine-related hospital admissions are thought to be preventable, which represents around 4% of all hospital admissions. Previous studies reported how often this happens for very specific and well-defined medicine regimes but primarily for elderly patients only.

What did you do? We looked at 18 medicine prescribing and 4 medicine monitoring indicators that help identify patients who might have been prescribed potentially harmful medicines or not being monitored properly after being prescribed certain medicines. We studied whether patients put at risk, according to the 22 indicators, had things in common and also whether the general practices at which these patients were registered shared common characteristics.

What did you find? We found that as patients get older and are prescribed more medicines they are at higher risk of potentially hazardous prescribing. However, they are at less risk of missed monitoring. Medication monitoring was also better in general practices that train doctors. Practices in poorer areas had higher chances of missing medication monitoring compared with practices in more affluent areas. We found no relationship between potential monitoring problems and potential prescribing errors.

What insights/knowledge did you add? The frequency with which some of the indicators occur were reported for the first time in this study. We showed, for the first time, that hospital care data plays a crucial role in medication safety surveillance. We analysed, for the first time, medication prescribing and monitoring as two different clinical processes and showed that patient and practice characteristics affect these processes in different ways.

Publication title: Developing a primary care patient measure of safety (PC PMOS): a modified Delphi process and face validity testing

Publication details (Vancouver format) Hernan AL, Giles SJ, O'Hara JK, Fuller J, Johnson JK, Dunbar JA. Developing a primary care patient measure of safety (PC PMOS): a modified Delphi process and face validity testing. BMJ Quality & Safety 2015;10.1136/bmjqs2015-004268.

What was known before your paper was published? Primary care is the first point of contact with the healthcare system for the majority of patients. But there is little research available that helps us understand how to improve safety. One way is to involve patients. Patients from hospital settings are often aware of issues around patient safety and have been able to identify when things go wrong. Patient feedback tools have been developed to capture what patients think contribute to errors in hospitals. This feedback has been used to reduce mistakes and improve care delivery. Developing a similar tool for primary care settings is needed.

What did you do? Using the hospital patient measure of safety (PMOS) tool as a guide, we adapted it for the primary care setting. To do this we used the existing survey questions from the hospital based PMOS survey, data from interviewed primary care patients about their views on safety, and the available literature to develop specific survey questions for the primary care setting. An expert panel representing academics, members of the public and health professionals agreed on what questions should be included. The survey was then pilot tested with patients and health professionals in Australia and the UK.

What did you find? The primary care patient measure of safety (PC PMOS) questionnaire contains 50 questions that cover 15 areas of safety. Some of these areas of safety include communication, access to care, information flow, organisation and care planning, equipment and environment, and staff training and education. The PC PMOS was assessed by a group of participants in a small pilot test, who believed that the PC PMOS was easy to understand and complete. Participants also believed patients would be willing to complete such a questionnaire with few barriers.

What insights/knowledge did you add? This is the first primary care patient measure of safety tool available. The PC PMOS is a quick and efficient way to get patientsâ&#x20AC;&#x2122; views on factors that contribute to patient safety in primary care. It helps to involve patients in improving safety and provides primary care organisations with practical information that can be used to prevent harm. At the end of the project, PC PMOS will be able to be used potentially as part of a thorough approach to safety management in primary care.

Publication title: Effectiveness of case management for ‘at risk’ patients in primary care: A systematic review and meta-analysis

Publication details (Vancouver format) Stokes J, Panagioti M, Alam R, Checkland K, Cheraghi-Sohi S, Bower P. Effectiveness of case management for ‘at risk' patients in primary care: A systematic review and meta-analysis. PLoS ONE. 2015; 10(7):e0132340.

What was known before your paper was published? Multiple chronic health conditions in a single person can cause patient safety concerns. One of the most notable is the concern of co-ordinating care provided by multiple practitioners in the health and social care services (e.g. between visits to a GP, specialist doctors in hospitals, and social workers). When information is not passed between providers efficiently and effectively, a patient may slip through the net, and end up with an emergency admission to hospital, which could otherwise have been prevented; this is bad for the patient in terms of health outcomes and the health system in terms of cost.

What did you do? One of the most popular interventions to try to avoid these breakdowns in co-ordination is called ‘case management of at-risk patients in primary care’, a form of so-called ‘integrated care’. This is where a patient is identified at increased risk of emergency admissions to hospital, and they are assigned a specific case manager/case management team to help deliver care specifically for that person and co-ordinate their care. While extremely popular, the actual effectiveness of this intervention is unclear. We looked at the existing global scientific literature and put together all of the relevant studies which assessed the effectiveness of the ‘case management of at risk patients in primary care’ intervention. We used a statistical technique, called ‘meta-analysis’, to put together the findings of all the studies we found, effectively re-creating the outcomes of one huge study of the intervention, which included all the participants of the smaller studies. This technique means we can be more confident in our findings. For that reason, it’s often referred to as the gold-standard form of evidence in medicine.

What did you find? We looked at effectiveness against a number of outcomes, using the goals that the health system tries to achieve: Health (self-assessed health status; mortality), cost-effectiveness (total cost of care; use of primary and non-specialist care e.g. GP, social workers; use of hospital care), and satisfaction (patient satisfaction). We found that case management of at-risk patients in primary care showed no significant effect on most of the outcomes we looked at. The exceptions were self-assessed health status (which may increase very slightly in the short, but not long-term), and patient satisfaction (which does appear to increase a small amount in both the short and long-term).

What insights/knowledge did you add? We showed the limited effectiveness of a popular method currently used by policy-makers and those who commission care in order to â&#x20AC;&#x2DC;integrate careâ&#x20AC;&#x2122;. This means that policy-makers need to broaden their thinking into what might work better to improve the health care system for those with multiple conditions. We suggest that aiming at a small proportion of the highest risk patients will only ever have limited effectiveness for the health system as a whole. Care should be better coordinated for all patients in the health service, and more efficient methods need to be explored, rather than relying solely on interventions with limited effectiveness. However, managing those high-risk patients is also important, and while the intervention analysed showed limited benefit, there was no harm shown. Therefore, we also tried to pick out some key learning opportunities for policy-makers to try and improve the interventions delivered in the meantime.

Publication title: Missed Diagnostic Opportunities and English general practice: a study to determine their incidence, confounding and contributing factors and potential impact on patients through retrospective review of electronic medical records.

Publication details (Vancouver format) Cheraghi-Sohi S, Singh H, Reeves D, Stocks J, Morris R, Esmail A, Campbell S, de Wet C. Missed Diagnostic Opportunities and English general practice: a study to determine their incidence, confounding and contributing factors and potential impact on patients through retrospective review of electronic medical records. Implement Sci. 2015 Jul 29;10:105. doi: 10.1186/s13012-015-0296-z. Erratum in: Implement Sci. 2015;10:124. PMID:26220545

What was known before your paper was published? Patient safety (the prevention of errors and adverse effects on patients associated with health care) has received considerable global attention since the publication of a landmark report ‘To Err is Human’ in 1999. However, the majority of research has focused on hospital settings despite the fact that in many countries the majority of patient contacts are in the community or primary care (general practice, pharmacies, dental practices etc). What we know about patient safety in primary care is developing but we still know little. Diagnostic error is a relatively understudied and an unmeasured area of patient safety, especially in general practice. Diagnostic error rates vary according to how ‘error’ is defined but one suggested criterion is clear evidence of ‘missed opportunity’ to make a correct or timely diagnosis to prevent them.

What did you do? While there is no agreed definition or method of measuring missed diagnostic opportunities, looking at patients’ medical records is the best approach. This study aims to: 1) find out how often missed diagnostic opportunities occur (the incidence) in English general practice; 2) identify what factors cause missed diagnostic opportunities; and 3) determine the (potential) impact of the detected missed diagnostic opportunities on patients. The research will consist of two stages, both of which will look at the electronic health records of patients in general practices in the Greater Manchester area of the UK. In the first phase we shall work in 15 general practices. Several general practitioners (GPs) are taking part as reviewers of medical records. In each practice, (continued)

two of the GPs will each look at the same sets of patient records on their own and we shall assess how well they agree in identifying and assessing missed diagnostic opportunities. The findings will enable us to make a preliminary estimate of how often missed diagnostic opportunities occur in general practice and this information will be used to calculate the number of records that will need to be reviewed in the second phase in order to estimate the true incidence of missed diagnostic opportunities in general practice. Up to 35 practices will take part in phase two. In each practice, a random selection of 100 patient records will be reviewed where a face-toface consultation occurred between 1 April 2013 and 31 March 2014 and the patient is aged 18 or older.

What did you find? Findings from the study will be published in a separate paper, once the data has been collected and analysed.

What insights/knowledge did you add? There are no reliable estimates of safety problems related to diagnosis in English general practice. This publication outlined the method that will be used to make an estimate of missed diagnostic opportunities. The results from this study will enable us to improve the safety of diagnoses in general practice and give us a better understanding of their impact on patients and what causes them to help us identify ways to reduce the chance of them occurring in the future.

Publication title: Blame the patient, blame the doctor or blame the system? A meta-synthesis of qualitative studies of patient safety in primary care

Publication details (Vancouver format) Daker-White G, Hays R, McSharry J et al. Blame the patient, blame the doctor or blame the system? A metasynthesis of qualitative studies of patient safety in primary care. PLOS One. 2015;pone.0128329

What was known before your paper was published? Most patient safety research has been conducted in hospitals. Recent years have seen more studies in primary care. We wanted to bring the findings of the primary care studies together to see if we could advance knowledge in the field and to help us examine findings from one of our own studies. To date, other studies have listed the kinds of errors or adverse events found in primary care. These typically involve errors related to communication, office or computer processes, treatments and clinical knowledge. Issues raised by patients are broader and include unmet expectations and violation of trust.

What did you do? We searched for published articles that reported qualitative studies about patient safety in primary care. A qualitative study is one where people are interviewed (alone or in groups) and/or observed by researchers who then analyze the words spoken in interviews or situations, rather than counting things or doing statistics. Once we had found articles, we assessed the quality of the reported research and tried different ways of putting the findings together until new insights were learned about patient safety in primary care (including pharmacy and general practice).

What did you find? Factors seen to improve or worsen safety for patients boiled down to the characteristics or behaviour of patients, staff or clinical systems and interactions between staff, patients and staff, or people and systems. Electronic health records, protocols and guidelines could be seen to both degrade and improve patient safety in different circumstances. For patients, â&#x20AC;&#x2DC;patient safetyâ&#x20AC;&#x2122; appears to be an individual feeling or judgement grounded in moral views and with potentially hidden psychological consequences affecting care processes and relationships. The main threats to safety appeared to derive from systems issues, for example involving service accessibility, resources or working hours.

What insights/knowledge did you add? Overall, the findings underline the human elements in patient safety primary health care. The key to patient safety lies in effective face-to-face communication between patients and health care staff or between the different staff involved in the care of an individual patient. Electronic systems can compromise safety when they override the opportunities for face-to-face communication. The circumstances under which guidelines or protocols are seen to either compromise or improve patient safety needs further investigation.

Publication title: Multimorbidity and Patient Safety Incidents in Primary Care: A Systematic Review and Meta-Analysis

Publication details (Vancouver format) Panagioti M, Stokes J, Esmail A, Coventry P, Cheraghi-Sohi S, Alam R, Bower P. Multimorbidity and Patient Safety Incidents in Primary Care: A Systematic Review and Meta-Analysis. PLoS One. 2015;10(8):e0135947.

What was known before your paper was published? Multimorbidity (a person with two or more long-term health conditions) is increasingly common. Primary care organisations (e.g. general practice) are increasingly responsible for the care of patients with long-term conditions, and improving the quality of their care is a major policy priority. Patient safety is a key component of high quality care, but has been a neglected research area to date. ‘Safety incidents’ can be a variety of things, but include anything that led (‘active safety incidents’) or could have led (‘precursors’) to unintended health consequences for the patient that results from what is done to, with or by the patient (processes of care), and not related to the patient’s underlying condition, e.g. adverse drug reaction from taking the wrong drug/dose. It is currently uncertain whether there are any specific groups of patients with multimorbidity who are more susceptible to patient safety incidents and whether there are any particular types of patient safety incidents that are more common in patients with multimorbidity.

What did you do? We searched the existing global scientific literature and put together all of the relevant studies which assessed the relationship of multimorbidity and patient safety incidents. We pulled together all of the findings (using statistical technique, called ‘meta-analysis’), , effectively re-creating what would have been the outcomes of one large study from the data from the participants of the individual smaller studies. This technique increases the accuracy of our analyses and means we can be more confident in our findings. For that reason, it’s often referred to as the gold-standard form of evidence in medicine. This review sought to answer two research questions: 1. Are patients with multimorbidity more vulnerable to patient safety incidents? 2. Does the relationship between multimorbidity and patient safety vary across different types of multimorbidity and different types of patient safety outcomes?

What did you find? The association between multimorbidity and patient safety is complex, and varies by type of multimorbidity and type of safety incident. Our analyses suggest that multimorbidity involving mental health may result in more patient safety incidents, which has important implication for the design and targeting of interventions to improve patient safety.

What insights/knowledge did you add? We highlighted the importance of mental health conditions in affecting the number of patient safety incidents in primary care for multimorbid patients. High quality studies examining the mechanisms of patient safety incidents in patients with multimorbidity are needed, with the goal of promoting effective service delivery and reducing threats to safety in this group of patients.

Publication title: The care.data concensus? A qualitative analysis of opinions expressed on Twitter

Publication details (Vancouver format) Hays R, Daker-White G. The care.data consensus? A qualitative analysis of opinions expressed on Twitter. BMC Public Health. 2015;15:838

What was known before your paper was published? At the start of 2014, NHS England began informing the public about the care.data project, whereby information would be collected from patientsâ&#x20AC;&#x2122; medical records and stored in a large database. The announcement sparked intense debate through traditional and social media, about how and why this project would be carried out. Discussions continued when, on 19 February 2014, NHS England announced the start of the project would be postponed. Many people shared information and opinions about care.data on Twitter, a social media website used to keep up-to-date with current events. Increasingly, Twitter is also being used by health services researchers as a source of data.

What did you do? Our aim was to identify and describe the range of opinions expressed on Twitter about care.data, around the time the delay to the project was announced. We searched for public tweets containing the hashtag #caredata, at regular intervals, until we had obtained approximately 10,000 tweets. All retweets (6,118, 62 %) and spam (240, 2 %) were excluded. The remaining 3,537 tweets had been posted by 904 contributors over 18 days during February and March 2014. We used methods of qualitative data analysis and read all tweets to identify emerging themes. Tweets were coded for each theme they contained, and analysed in-depth within and between themes.

What did you find? Tweets containing the hashtag #caredata could be grouped into one or more of 9 key themes: 1. 2. 3. 4. 5. 6. 7. 8. 9.

Informed consent and the default ‘opt-in’ as opposed to ‘opt-out’. Lack of trust. Concerns about privacy and data security. Data management and access by private companies. Legal issues, data protection, and GPs’ concerns. Communication failures and confusion about care.data. Reasons for the delayed implementation. Lack of patient involvement and patient-centeredness. The potential benefits of care.data and the ideal model of implementation.

What insights/knowledge did you add? Using methods of qualitative data analysis enabled us to identify and understand a range of concerns about the care.data project, and how these might be overcome. We discovered those for and against the project shared many concerns, and our findings highlighted the risks of not considering and addressing public opinion. For example, we identified the potential for patient safety incidents resulting from a lack of trust in the healthcare system. Tweets can be used to provide insight about peoples’ views on a topic. However, those contributing to discussions on social media may lie more heavily on one-side of a debate than another. Thus, we do not assume our findings reflect the majority public opinion.

Publication title: Examining variations in prescribing safety in UK general practice: cross sectional study using the Clinical Practice Research Datalink

Publication details (Vancouver format) S Jill Stocks, Evangelos Kontopantelis, Artur Akbarov, Sarah Rodgers, Anthony J Avery, Darren M Ashcroft. Examining variations in prescribing safety in UK general practice: cross sectional study using the Clinical Practice Research Datalink. BMJ 2015;351:h5501 doi: 10.1136/bmj.h5501

What was known before your paper was published? â&#x20AC;&#x2DC;Prescribing safety indicatorsâ&#x20AC;&#x2122; describe patterns of prescribing that should not usually happen, for example somebody who has had a stomach ulcer in the past should not usually be prescribed a non-steroidal antiinflammatory drug (NSAID) such as ibuprofen. A group of GPs had agreed previously on the definition of several prescribing safety indicators and a trial found that a pharmacist working in general practices could reduce how often some types of potentially hazardous prescribing occurred. Some work had been done in Scotland to see how often potentially hazardous prescribing occurred in general practice using similar but not identical prescribing safety indicators, but it had not been measured in general practices across all of the UK.

What did you do? Our research question was: What is the prevalence of different types of potentially hazardous prescribing in UK general practice, and what is the variation between practices? We used anonymised electronic health records from 526 general practices that are part of a large research database called the Clinical Practice Research Datalink. From this database we identified all adult patients who were exposed to the possibility of being given a prescription that they usually should not have (because of a diagnosis or prescription they already had). We can think of this as being in a position to trigger the prescribing (continued)

safety indicator. Then we counted how many of these patients were actually given a prescription that they usually should not have. We can think of this as triggering the prescribing safety indicator. Our results were the proportion of patients who triggered a prescribing safety indicator.

What did you find? We found that some prescribing safety indicators were quite common in some general practices, meaning that potentially hazardous prescribing does occur across the UK. There were large differences between practices in how often prescribing safety indicators occurred and this was still true even if we considered other factors that might explain why there might be differences between practices, such as size or location of the practice. Older patients and those taking more repeated medications were more likely to trigger a prescribing safety indicator, whilst younger patients taking fewer repeated medications were more likely to not have a blood monitoring test within the recommended time period.

What insights/knowledge did you add? The high variation between practices suggests that we might be able to reduce how often prescribing safety indicators are triggered by making improvements at the practice level. We also emphasise that older patients and those taking multiple medications have the highest risk and regular medication reviews are especially important for this type of patient.

Publication title: Sensemaking and the co-production of safety: a qualitative study of primary medical care patients

Publication details (Vancouver format) Rhodes P, McDonald R, Campbell S, Daker-White G, Sanders C. Sensemaking and the co-production of safety: a qualitative study of primary medical care patients. Sociol Health Illn. 2015 Nov 6. doi: 10.1111/1467-9566.12368. [Epub ahead of print]

What was known before your paper was published? Research looking at patientsâ&#x20AC;&#x2122; understandings of safety in primary care has mostly looked at how patients define errors in their care. It has found that the range of patient reported harms and their causes tends to be wider than what is generally recognised by healthcare professionals. This indicates that patient safety is broader than the absence of error. Recent research has shown that safety can be health professionals interacting with each other as well as the established importance of patients and their friends and family have in healthcare safety. However, patientsâ&#x20AC;&#x2122; perspectives and experiences of safety are likely to vary in different situations. Most existing work has looked at hospital or hospice settings, rather than primary care, even though most contacts with healthcare are in primary care (general practice, pharmacy, dentistry etc). Sensemaking is a concept developed by Weick (1995) as a way of viewing healthcare professional behaviours and actions in relation to patient safety. The concept of sensemaking describes the process by which people interact with their environment and experiences.

What did you do? We asked individuals how they make sense of their experiences of primary healthcare and how sensemaking influences their views of safety. We wanted to find out also whether the findings in primary care agree with those in hospital settings (Doherty and Saunders 2103) in terms of how primary care patients engage in sensemaking and the concept of safety as something that is co-produced by health professionals and patients.

What did you find? We found three core themes: (i) trust and psycho-social aspects of professional-patient relationships (perceptions about whether patients feel their GP trusts them and whether they trust their GP) (ii) choice, continuity, access (the patient being able to see the doctor they want or not being able to for whatever reason, and the advantages and disadvantages of both) (iii) organisational and system-level tensions constraining safety (e.g. lost test results or inaccurate prescriptions). The main findings in relation to these core themes have been reported elsewhere (Rhodes et al. 2015; Rhodes et al. 2014). However, for this paper, we re-analysed the data using Weickâ&#x20AC;&#x2122;s framework to understand sensemaking (behaviours and actions in relation to patient safety) amongst primary care patients. We found that what makes patients feel safe is often very different from the sorts of things that policy makers and healthcare professionals focus on. Patientsâ&#x20AC;&#x2122; accounts suggested that safety is not something that can be easily identified and measured. Rather, in many cases, safety emerges from the interaction between people, and between people and their environment. Patients found it much easier to draw on personal experience when describing and making sense of safety.

What insights/knowledge did you add? This study explored the ways in which patients make sense of 'safety' in the context of primary care. Using interviews with primary care patients, we found that patients' views of safety are dependent on the situation in which a patient experiences a potential safety incident. Patients reported taking action to protect themselves from potential harm. This demonstrated that patients were not passive recipients of care. Instead they used their knowledge and experience to inform their actions. The study highlighted the differences and similarities between hospital and primary care settings. It also concluded that a broad definition of patient safety is required by policy makers, health professionals and patients which includes the safety concerns of patients in primary care settings.

Publication title: Non-psychotropic medication and risk of suicide or attempted suicide: a systematic review

Publication details (Vancouver format) Gorton HC, Webb RT, Kapur N, Ashcroft DM. Non-psychotropic medication and risk of suicide or attempted suicide: a systematic review. BMJ Open 2016;6:e009074 doi:10.1136/bmjopen-2015-009074

What was known before your paper was published? It is known that people diagnosed with most mental illnesses have higher risk of suicide than those who do not have these illnesses. Certain physical illnesses may also be linked with increased suicide risk. There have been occasional reports that some medicines used to treat physical illness, ‘non-psychotropic medication’, might increase risk of suicide. It is difficult to measure suicide risk in clinical trials because these relatively small studies cannot detect these tragic but rare events. Studies which use real-life healthcare data, known as ‘observational studies’, can help our understanding of factors associated with suicide risk. There are no comprehensive literature reviews examining the risk of suicide or attempted suicide that may be linked with these medications.

What did you do? We aimed to find out: a) which non-psychotropic medications have already been investigated b) whether they are suggested to change risk of suicide or attempted suicide and c) evaluate the strengths and limitations of the study designs used. We searched four large electronic literature databases and found 19 studies which were eligible for inclusion. We identified the medication groups which were investigated by each study and whether they had been associated with an increased or decreased risk of suicide or attempted suicide, or no change in risk at all. We then carefully reviewed the study for strengths and weaknesses which might have influenced the findings each study generated.

What did you find? The following groups of medicines have been investigated in relation to risk of suicide and attempted suicide: 1) drugs used to treat epilepsy 2) drugs used for heart disease 3) steroids 4) asthma medications called leukotriene receptor antagonists 5) a drug used to help people stop smoking called varenicline 6) a drug to treat severe acne called isotretinoin and 7) a particular antibiotic group called quinolones. It was difficult to compare results between the studies because they all had different designs and accounted for other risk factors for suicide to different extents. Drugs for heart disease did not seem to increase risk of suicide, but we could not tell whether the other medication groups influenced suicide risk or not from the available evidence.

What insights/knowledge did you add? We found many difficulties when trying to compare risk of suicide and attempted suicide between different types of non-psychotropic medication. Therefore, we are not confident to say that these medication groups do or do not influence suicide risk. To understand this further, additional better-designed studies are needed. Ideally, future studies would be sufficiently large and robust in their design to enable the investigators to unpick whether any heightened risk of suicide observed was due to medication prescribed or due to other risk factors the patient has, especially underlying mental and physical illnesses.

Publication title: Effectiveness of multidisciplinary team case management: difference-in-differences analysis

Publication details (Vancouver format) Stokes J, Kristensen SR, Checkland K, Bower P. Effectiveness of multidisciplinary team case management: difference-in-differences analysis. BMJ Open. 2016; In print

What was known before your paper was published? ‘Case management’ is where a ‘high-risk’ (usually at risk for future hospital admissions) patient is: • • • •

identified their individual needs are assessed by a team or a single case manager an individualised care plan is prepared, and care is co-ordinated the care plan is regularly reviewed, monitored and updated.

The assumption is that this greater individual attention and co-ordination effort by the healthcare professionals in primary care will prevent any future unnecessary hospital admissions. We previously undertook a systematic literature review where we found little effect of case management for high risk patients directly treated. The review found little effect in terms of health status, use of hospital care, costs of care, but a small increase in patient satisfaction. We thought that case management conducted by a multidisciplinary team composed of various healthcare professionals and staff, including a social worker, may be slightly more effective. However, we identified some gaps in the literature, including a lack of studies in ‘real world’ settings i.e. not in a controlled experimental setting, but where the intervention has already been rolled out to the population it is supposed to help. There was also little analysis of potential ‘spill-over effects’ where all the patients of that particular general practice may benefit from better team working, or increased ‘professional integration’, rather than just those individual patients who received specific case management.

What did you do? We looked at a local integrated care intervention with a multidisciplinary team (including a GP, practice nurse, district nurse, social worker, active case manager, and access to specialists outside of the core team) offering

(What did you do? continued… ) case management to those considered to be in the top 2% highest risk for hospital admissions of the practice population. We used a technique called difference-in-differences analysis (a method used in economics research, that can get as close as possible to eliminating bias from our results, and can be used in a ‘real world’, instead of experimental setting). We measured a variety of hospital use and cost measures at both the individual level (to see the effect of integrated care on those directly treated through case management compared to patients who were not given specific case management), and at the practice level where the intervention was rolled out gradually to practices, so we were able to measure indications of ‘spill-over effects’.

What did you find? Similar to the results of our systematic review, we found no clinically significant effects of case management for high risk patients, even with a multidisciplinary team and social worker involved in the case management. For those highest risk patients treated (according to risk tool score), some measures of hospital use were very slightly increased by the intervention, so the highest risk patients treated appeared to benefit least. At the practice level, we identified indications of a beneficial spill-over effect, with a small reduction in inpatient emergency admissions. However, the statistical significance of this result was questionable.

What insights/knowledge did you add? We have added to the evidence base for an extremely widespread intervention for ‘integrating care’ with a ‘real world’ study looking at effects for those patients directly treated, as well as other patients within the same practice. At the individual patient level, we saw similar findings to our systematic review. Our previous thought that there may be added benefits for the patient to be treated by a multidisciplinary team including a social worker did not stand up. The evidence from the previous review was mostly from a ‘low strength’ primary care system, the USA, and perhaps the effects of the case management intervention are more exaggerated in a system like that, where case management may essentially substitute for ‘good primary care’. We found that the highest risk patients appeared to benefit least, so the benefits of focusing efforts on system changes for this group of patients are questionable. Our findings of a possible spill-over effect require further confirmation, but the possibility of this effect raises interesting issues. We would like to see some qualitative/process evaluation work to further explore these findings. Our next piece of work will attempt to identify any ‘multimorbidity’ subgroups that may benefit more directly from an ‘integrating care’ intervention.

Publication title: When procedures meet practice in community pharmacies: qualitative insights from pharmacists and pharmacy support staff

Publication details (Vancouver format) Thomas CEL, Phipps DL, Ashcroft DM. When procedures meet practice in community pharmacies: qualitative insights from pharmacists and pharmacy support staff. BMJ Open. 2016;6(6): e010851.

What was known before your paper was published? Procedures are official or established ways of doing things. They are used in many industries such as rail and aviation, oil and gas and healthcare settings. We knew that standardised procedures are a common method for trying to ensure that staff members work safely, however the presence of procedures does not always guarantee that they are followed. Previous research has shown that sometimes staff will deviate from procedures as part of their work for a range of reasons. We wanted to know more about how this happens in community pharmacy settings as up until now there has been very little research into procedural compliance in this setting.

What did you do? We interviewed 13 pharmacists and 11 pharmacy support staff from a range of different community pharmacies. During the interview, we asked participants a range of questions to understand how procedures influenced their work. The interview included questions such as ‘what kind of procedures do you have to follow in community pharmacy?’, ‘how useful are the procedures for helping you to do your job?’ and ‘how do you feel procedures benefit you in your work?’

What did you find? We found that community pharmacy staff members are expected to follow a large number of procedures as part of their work. At times, staff felt that it was impossible to follow all procedures fully. There were many factors that at times made following procedures difficult, including a high workload, a lack of staff, poor communication and a pressure to hit targets. Staff also spoke about bypassing or deviating from procedures to ensure that tailored patient care was provided. This study supports the need for pharmacists to be able to use their professional judgement when deciding if it is appropriate to bypass or deviate from procedures for patient safety.

What insights/knowledge did you add? We found that community pharmacy staff members display â&#x20AC;&#x2DC;organisational resilienceâ&#x20AC;&#x2122; (the ability to react to changing situations accordingly in order to maintain high performance). We also highlighted the impact of factors such as a high workload and a lack of staff can have on the ability of staff to follow procedures. These findings should help to inform policy makers about how they can best support staff in their work and in maintaining safe and effective patient care. The paper has been published in BMJ Open and is freely available to read: http://bmjopenbeta.bmj.com/content/6/6/e010851

Publication title: An external validation of models to predict the onset of chronic kidney disease using population-based electronic health records from Salford, UK

Publication details (Vancouver format) Fraccaro P, van der Veer S, Brown B, Prosperi M, Oâ&#x20AC;&#x2122;Donoghue D, Collins GS, et al. An external validation of models to predict the onset of chronic kidney disease using population-based electronic health records from Salford, UK. BMC Med [Internet]. 2016;14(1):1â&#x20AC;&#x201C;15.

What was known before your paper was published? The number of patients diagnosed with chronic kidney disease (CKD) is increasing worldwide. Identifying people who may develop CKD would enable early preventative treatment that would reduce the impact of this disease. Several ways of identifying people at increased risk of developing CKD have been proposed in the scientific literature using risk models. However, to date only a few of them have been tested and directly compared. We were particularly interested in which one of these risk score models may be used in English primary care.

What did you do? We selected CKD risk score models from the scientific literature that could be used in the English primary care system and tested their performance in identifying people at high risk of developing CKD within 5 years. We used linked, anonymised, structured primary and secondary healthcare data from patients resident in Salford, UK. We included in our study all patients who had any contact with their GP in 2009 and followed them until the end of 2014, death, or diagnosis of CKD. For each risk score model, we assessed whether the model could discriminate people who did develop CKD compared to those people who did not, and whether the model was well calibrated for the English population.

What did you find? We found seven CKD risk score models that could be used in English primary care. All the models were able to discriminate between people who developed CKD and those people who did not. However, most of the models were a poor match to our English study population, with the models identifying people as high risk although they did not develop CKD within our study period. Only two models were calibrated to accurately identify people at high risk in our study population.

What insights/knowledge did you add? The CKD risk score models we evaluated showed good discriminative ability but were poor at identifying people at high risk of developing CKD who were English primary care patients. QKidney, the only UK-developed model, outperformed the others, and could be conceivably used in the English primary care to identify people who are at high risk and therefore receive early preventative treatment before CKD develops.

Publication title: Integrated care: theory to practice

Publication details (Vancouver format) Stokes J, Checkland K, Kristensen SR. Integrated care: theory to practice. Journal of Health Services Research & Policy. 2016, Vol. 21(4) 282–285 DOI: 10.1177/1355819616660581

Summary of publication Health systems internationally face a common set of challenges: ageing populations, increasing numbers of patients suffering from multiple long-term conditions (multi-morbidity) and severe pressure on health and care budgets. ‘Integrated care’ is pitched as the solution to current health system challenges. But, in the literature, what integrated care actually involves is complex and contested. Multi-disciplinary team case management (involving individualised care planning by a number of different professionals attempting to consider the multiple needs of the individual patient) is frequently the primary focus of integrated care when implemented internationally. In this perspective article, we examine the practical application of integrated care in the NHS in England to exemplify the prevalence of the case management focus. We look at the evidence for effectiveness of multi-disciplinary team case management, for the focus on high-risk groups and for integrated care more generally. We suggest realistic expectations of what integration of care alone can achieve and additional research questions.

Publication title: Temporal trends in antidepressant prescribing to children in UK primary care, 2000-2015

Publication details (Vancouver format) Sarginson J, Webb RT, Stocks SJ, Esmail A, Garg S, Ashcroft DM. Temporal trends in antidepressant prescribing to children in UK primary care, 2000-2015. J Affect Disord. 2017;210:312-318.

What was known before your paper was published? During the 1990â&#x20AC;&#x2122;s a steady increase in antidepressant prescribing for children and adolescents was observed. In the early 2000â&#x20AC;&#x2122;s reports of an increased risk of suicidal thoughts and behaviour sparked safety concerns and a sharp reversal in antidepressant prescribing in this age group. Despite these widely publicised warnings and the introduction of new guidelines in 2005 for treatment of depression in children, antidepressant prescribing in under 19s increased sharply between 2005 and 2012 in the US, UK and other European countries. However, relatively little is known about why these antidepressants were being prescribed and to whom, and whether this trend continued after 2012.

What did you do? In order to see if antidepressant prescribing in children has continued to increase since 2012 we looked at changes in the frequency at which antidepressants were prescribed to 5-17 year olds by GPs between 2000 and 2015. This was done using the anonymous primary care patient records for approximately 6.8% of the total UK population. Additionally we looked at which groups of children were most likely to be prescribed an antidepressant and what was the most likely reason for the prescription.

What did you find? By 2014 antidepressant prescribing levels in 5-17 year olds were higher than those seen prior to 2005. In 2015 approximately 3 in every thousand 5-17 year olds with no previous history of antidepressant use were prescribed an antidepressant for the first time. Sixty one percent of these prescriptions were to 15-17 year old females. For those prescribed an anti-depressant for the first time, only 21% of the prescriptions were linked to a diagnosis of depression, 13% to symptoms of depression and 20% of prescriptions linked to alternative indications with relevant guidelines for their use in children. These alternative indications included the treatment of anxiety, bulimia nervosa, obsessive-compulsive disorder, chronic pain and migraines. The final 46% of prescriptions issued either could not be linked to a diagnosis or were linked to alternative use without relevant guidelines for their use in children.

What insights/knowledge did you add? During 2015, not all of the children and adolescents receiving an antidepressant prescription for the first time had a diagnosis linked to depression or an alternative use with age appropriate guidelines. Also 82% of these antidepressants prescriptions were issued to adolescents aged 15-17. This is the age range during which most young people start to transition from child and adolescent mental health service to adult services. This can make them vulnerable to disruptions in treatment or present additional difficulties for those trying to identify and access appropriate services for young people.

Publication title: Understanding the implementation and adoption of a technological intervention to improve medication safety in primary care: a realist evaluation

Publication details (Vancouver format) Jeffries, M., Phipps, D. L., Howard, R. L., Avery, A. J., Rodgers, S., & Ashcroft, D. M. (2017). Understanding the implementation and adoption of a technological intervention to improve medication safety in primary care: a realist evaluation. BMC Health Services Research, 17(1), 196. DOI: 10.1186/s12913-017-2131-5.

What was known before your paper was published? Over a billion prescription items are issued each year in the UK. It is known that there are risks associated with the prescribing of medicines in primary care and that for some patients this can lead to an adverse drug event (ADE). An estimated 9-12% of hospital admissions are caused by ADEs. Using information technology (IT), such as electronic medicines optimisation systems, can be used by doctors, pharmacists and healthcare managers to gain easier access to prescribing data. This can allow clinicians to make assessments about the quality and safety of prescribing and for changes to be made to ensure medicines are prescribed more safely. However the introduction of such IT systems has previously encountered difficulties. It has been acknowledged that the introduction of IT systems in to healthcare settings needs to consider the people who will be using the technology, the ways people work and the context in which the technology is used.

What did you do? We talked to members of a clinical commissioning group in the South of England and looked at the use of an IT system designed to support medicines safety in primary care. This web based system securely extracted patient data from general practice records. We talked to GPs, general practice managers, pharmacists and patients, in interviews and focus groups, to explore their experiences of working with the IT system and what they felt were the benefits and drawbacks of it. We used a â&#x20AC;&#x153;realist evaluationâ&#x20AC;? perspective which seeks to explain the ways an intervention might work (in this case the IT system), for whom and under what circumstances.

What did you find? Using the IT system could lead to a number of improved patient safety outcomes, when the system was used effectively. Patients at risk of adverse drug events could be reviewed and changes made to their medication. However, using the IT system effectively depended upon how different healthcare professionals used the system, the flow of information between different healthcare professionals, and how different ways of working were developed to make best use of the system. The use of the IT system was undermined by people thinking that they lacked ownership, by restricted access and by a poor knowledge and awareness of its benefits.

What insights/knowledge did you add? The use of an electronic medicines optimisation IT system may improve medication safety in primary care settings by identifying those patients at risk of an adverse drug event. To fully realise the potential benefits for medication safety there needs to be more usage of this or other similar IT systems across primary care and by a wider range of stakeholders. Engaging with all potential stakeholders and users prior to implementation of such IT systems might allay perceptions that the system is owned by the CCG and increase knowledge of the potential benefits, therefore leading to increased use of these types of IT systems in primary care.

Publication title: Antipsychotic prescribing to patients diagnosed with dementia without a diagnosis of psychosis in the context of national guidance and drug safety warnings: longitudinal study in UK general practice

Publication details (Vancouver format) Stocks SJ, Kontopantelis E, Webb RT, Avery AJ, Burns A, Ashcroft DM. Antipsychotic prescribing to patients diagnosed with dementia without a diagnosis of psychosis in the context of national guidance and drug safety warnings: longitudinal study in UK general practice. Drug Safety 2017;

What was known before your paper was published? Patients with dementia are sometimes prescribed antipsychotic drugs to help with the behavioural and psychological symptoms of dementia. However, according to NICE guidance, this should only be considered for severe symptoms and after other approaches have failed because of the increased risk of several health problems, particularly stroke. Before these health risks emerged, in the late 1990s and early 2000s, prescribing antipsychotic drugs to patients with dementia was quite common. Since then, several drug safety warnings and guidelines were published about these risks, between 2004 and 2012.

What did you do? We wanted to see if the drug safety warnings and guidance had worked. To achieve this we examined if the prescribing of antipsychotic drugs to patients with dementia had reduced after warnings and guidance were released, by analysing the anonymised primary care records of over 111,000 older patients with dementia.

What did you find? Antipsychotic drugs fall into 2 different groups, and the results varied across groups. For group 1 (older, firstgeneration antipsychotic drugs), there was a large reduction in prescribing from 8.9% in 2001 to 1.4% in 2014 but for group 2 (newer, second-generation antipsychotic drugs) there was hardly any difference between 2001 and 2014. However, even though there was little change for the group 2 antipsychotic drugs over the longer time period from 2001 to 2014, there were short term changes in prescribing in line with the drug safety warnings.

The publication of guidelines by the National Institute for Health and Care Excellence (NICE) had no immediate impact on prescribing but approximately 1 to 2 years later a longer term reduction commenced. Between 2012 and 2015, the reducing trend for prescribing of group 2 antipsychotic drugs ended. The likelihood that a patient with dementia would be prescribed an antipsychotic drug varied widely between different general practices.

What insights/knowledge did you add? A further, carefully worded, warning may be needed to reduce the longer-term prescribing of some types of antipsychotic drugs that does not result in an increase in the prescribing of other types of drugs, which may also be regarded as inappropriate. Given the high variation in the likelihood between practices of a patient being prescribed an antipsychotic drug, perhaps the numerous computer systems in UK primary care could be used to alert the GP when this is happening. Another option might be medication reviews by clinical pharmacists working in general practices. In general the impact of drug safety warnings and guidelines depends on multiple factors that are rarely investigated. Increased surveillance of the effectiveness of drug safety warnings and guidance is needed to improve the impact of future warnings and guidelines.

Publication title: Never Events in UK general practice: a survey of the views of General Practitioners on their frequency and acceptability as a safety improvement approach

Publication details (Vancouver format) Stocks SJ, Alam R, Bowie P, Campbell S, de Wet C, Esmail A, Cheraghi-Sohi S. Never Events in UK general practice: a survey of the views of General Practitioners on their frequency and acceptability as a safety improvement approach. J Patient Saf 2017;00: 00–00

What was known before your paper was published? In hospitals ‘Never Events’ are rare but serious preventable incidents, such as operations where something went wrong (e.g. wrong limb amputated). They must be reported to NHS England and information about how often Never Events are reported is regularly published to encourage learning within the NHS. This Never Events approach has not been used in general practice to date. A list of ten Never Events suitable for use in general practice has been developed through discussions with general practitioners (GPs) and other practice staff. However, we did not know how often these Never Events were happening, or their acceptability to GPs as a method to improve patient safety in general practice. A list of these ten Never Events is at the end of this document.

What did you do? We asked GPs in Scotland and Greater Manchester to anonymously estimate how often the ten Never Events had occurred in their practice, how often they thought they might occur in the future, how they had reacted to a Never Event happening and whether they agreed that the event should be called a ‘Never Event’. Altogether 556 GPs from 412 practices responded to our survey.

What did you find? Some Never Events happened very rarely, for example, prescribing aspirin to a child younger than twelve years occurred in less than 1% of practices during one year. Other Never Events had happened quite often, for example an abnormal investigation result being received by a practice but not reviewed by a clinician was reported in around half of the practices during one year. Most GPs thought that all ten of the Never Events could

happen within their practice during the next 5 years. Almost all GPs reported that they would take the occurrence of a Never Event extremely seriously and undertake a serious event analysis, a process used by practices to discuss and learn from mistakes or problems. The majority of GPs agreed that the Never Event title was appropriate for most of the Never Events but GPs who reported a Never Event to have occurred in their practice were less likely to agree with the Never Event title.

What insights/knowledge did you add? The Never Events approach does not transfer easily from hospital care to general practice suggesting that the more common Never Events might be useful to monitor safety in general practices and the rarer Never Events could be used to monitor patient safety in a similar way as in hospitals. GPs would need to be reassured that the use of Never Events should be to identify weaknesses in the system and prevent Never Events from occurring rather than as a way to criticise general practice. Replacing the “Never Events” title with one that describes the missed opportunity to prevent the incident might improve their acceptability to GPs.

List of 10 Never Events 1. Prescribing Aspirin for a patient less than 12 years old (unless recommended by a specialist for specific clinical conditions e.g. Kawasaki’s disease) 2. Prescribing Methotrexate daily rather than weekly (unless initiated by a specialist for a specific clinical condition e.g. leukaemia) 3. Adrenaline/Epinephrine is NOT available within minutes when clinically indicated for a medical emergency in the practice or GP home visit 4. Prescribing a teratogenic drug to a patient the clinician knows to be pregnant (unless advised to do so by a clinical specialist) 5. Prescribing systemic oestrogen-only Hormone Replacement Therapy for a patient with an intact uterus 6. A planned referral of a patient, prompted by clinical suspicion of cancer, is not sent 7. Ambulance transport is not arranged if this had been agreed when deciding to admit a patient as an emergency 8. A needle-stick injury due to a failure to dispose of ‘sharps’ in compliance with national guidance and regulations 9. Prescribing a drug to a patient that has correctly been recorded in the practice system as having previously caused her/him a severe adverse reaction 10. An abnormal investigation result is received by a practice but is not reviewed by a clinician

The Greater Manchester Primary Care Patient Safety Translational Research Centre was funded by the National Institute for Health Research (NIHR) for five years from August 2012. It was a partnership between The University of Manchester and Salford Royal NHS Foundation Trust.