Sharing a written medical summary with patients on the post-admission ward round: A qualitative study of clinician and patient experience

Rationale, Aims and Objectives: Sharing aspects of the traditional medical record with patients has been successful in primary and antenatal care, but has not been investigated in the UK inpatient setting. Our aim was to evaluate the impact on patient and clinician experience of providing patients with a written lay summary of their care-plan in the acute care setting. Method: We carried out a qualitative interview study on two acute medicine wards in an NHS University Teaching Hospital for a 4-week period in 2019. A summary record, designed in response to suggestions from doctors and patients from a previous study, was distributed to patients on the first ward round after admission. Eligible participants included all doctors and nurses working on and all patients and their families attending the acute medical units; patients were excluded if they lacked capacity to consent or were under 18. We interviewed 20 patients, 10 relatives, 10 doctors and 7 nurses. Conclusions: A summary record could empower patients and improve patient-doctor communication but would require additional clinician and administrative time.


| INTRODUCTION
There is minimal transparency in the writing of hospital medical records: patients rarely request to see them and what can be seen is hard to interpret. Verbal communication, the mainstay of the patientdoctor interaction, can be misinterpreted or forgotten 1 ; in a recent study in a UK hospital, 27% of patients did not remember being told their diagnosis. 2 Patients often feel disempowered and struggle to become involved in their care. 3 To enhance patient autonomy, patients need to be supported in making their own informed decisions. One proposed way of doing this is to increase patient access to their medical records.
The sharing of patient records has been shown to empower patients, improve medication adherence 4 and enhance patient satisfaction in primary care, [5][6][7] antenatal care, 8 and for specific diseases. 9,10 To our knowledge, however, written patient information has never been shared in the acute setting in the UK. Access to inpatient medical records in a U.S. hospital was found to improve patient understanding and provide a sense of control, whilst most clinicians felt that it did not affect their practice. 11 However, several concerns were raised, in particular that that patients could be overwhelmed with information and become more confused. 11 Previously stated anxieties include clinicians censoring records to remove unlikely but worrying differential diagnoses, difficult discussions such as end of life care, or medical uncertainty highlighted in medical notes to be explored by other colleagues. 12 There are questions about the security of these written notes and the risk of them being lost. 12 There is also the possibility of third parties, such as relatives, accessing patient information without consent. 12 An evaluation of patient and clinician views on sharing the medical record, and alternative approaches to improving communication, was conducted via in-depth interviews and a questionnaire which we have previously reported. 2,13 Participants were very positive about sharing a summary note: it would allow patients to review information about their care in their own time and support them in asking questions, without changing the clinical notes used by the treating team.
Such a note might alter the nature and quantity of communication between patients, doctors and nurses. 14,15 Clinician and patient participants made suggestions of the ideal content of such a note. The aims of this study were to explore (a) the impact on health care professionals, patients and their relatives and (b) the practical issues of sharing such a summary record with patients admitted to the acute medicine ward.

| METHODS
This was a qualitative interview study conducted in January 2019 in a University Teaching Hospital which uses electronic patient records.

| Development of written summary plan
A summary care pro forma (Appendix S1) was informed by the findings of a previous mixed methods study which asked patients and doctors what they would want included in a summary care record. 2 This included information about the patient's likely diagnosis, differential diagnoses, treatments, tests and space to write questions.
The University Teaching Hospital where this study took place uses an electronic patient records system which allows creation of 'dotphrases'. A 'dotphrase' is a section of predefined text that can be added to the medical notes with spaces for the clinician to fill in information for each patient. During the first ward round after admission (known as the post-take ward round, PTWR), the medical team added a summary plan 'dotphrase' at the end of the patient notes (Appendix S1). This was printed off by a ward clerk and handed directly to the patient, along with a notice that the summary was the patient's own responsibility (Appendix S2). If the medical team felt that providing a written summary plan would distress the patient, they could choose not to do so, and we enquired about such cases in subsequent interviews to determine frequency and reasons for this. The summaries were only written for the PTWR, not for further ward rounds. The process of creating and the content of the summaries were observed by members of the research team.

| Inclusion criteria
Patients were eligible for inclusion if they were admitted to one of the wards where the study was taking place and were given a written PTWR summary.
Doctors and nurses were eligible for inclusion if they worked on one of the wards where the study was taking place.

| Exclusion criteria
Patients who were aged less than 18, had cognitive impairment sufficient that they lacked capacity to consent or insufficient spoken English to undertake the consent process were excluded. Patients who had been given a written PTWR summary were identified by a member of the clinical team or a GCP-trained medical student and given a PIS on the ward. After they were given time to read it, a member of the research team answered any questions and took their consent. The consent form was signed immediately before the interview took place.

| Data collection and analysis
A semi-structured interview guide (Appendix S3) was developed based on data from patient and clinician interviews from a previous study, purposively sampled to represent a range of ages, clinician experience and gender. 2 Interviews were conducted by AF and AC, two GCP-trained medical students who were male and female respectively. Interviews with patients took place at the bedside or in a private room where available. Interviews with clinicians took place in their offices or a private room. Interviews lasted around 15 min and took place with two researchers: one writing down the interviewee statements verbatim in real time, while the other conducted the interview. Where participants wished to discuss things beyond the initial guide, this was annotated in free text; subsequent interviews were then adapted to include themes which emerged from initial interviews with patients, relatives and clinicians. Following each interview, the interviewers reviewed the notes to minimize risk of misinterpretation.
No repeat interviews were carried out. Interviews were conducted until data saturation was reached.
Interviews were double-coded using NVivo 12.5.0. Thematic analysis of the data was carried out with themes generated both deductively using the interview questions (see Appendix S3) and inductively from the data itself (A.F., C.L.). Coding differences were rare: where present, they were noted, discussed and a consensus was reached. Participants were not asked to provide feedback on the findings.

| RESULTS
Thirty-eight interviews were conducted in total: 10 doctors, 7 nurses, 13 patients and 1 relative were interviewed individually; 7 interviews were conducted with patients and their relative(s) together; in these cases, patient and relative views were analysed separately, leading to a total of 45 interview files. 13/33 patients approached chose not to be interviewed or were unable to be interviewed for reasons such as early discharge or not being at the bedside. All doctors and nurses approached agreed to be interviewed. Doctors interviewed reported that the only reasons for not giving summary notes to patients was because the patient lacked capacity, were confused, or, in one case, had visual problems.
The demographics of the patients and relatives interviewed were not recorded. Hospital-wide data on adult admissions for this time period showed 53.1% of patients were female (average age of 54.8) and 46.9% were male (average age of 59.9). The district council in which the hospital was based had an index of multiple deprivation (IMD) 2015 score of 13.8 and in 2019 the local authority district was ranked 205th out of 317 local authorities from most deprived to least deprived. 16 The IMD is the official measure of relative deprivation for small areas in England. In addition, 82.8% of residents listed English as their first language. 17 Illustrative quotes have the interviewer type identified by N = nurse; D = doctor; P = patient; R = relative.

| The summary empowered patients
The summary notes helped patients to better understand their care, through both improved recall and enhanced communication.  One perspective that was expressed several times was that having more information (such as seeing all their records) could cause worry through a lack of understanding or not having the information ade-

| Barriers to implementation
Questions about confidentiality were included in the interview template as we thought this could be a concern for patients. However, 20/28 patient and relative interviews specifically mentioned having no concerns about confidentiality: I see the worth in it so I don't mind… [it's] not information that anyone can do anything with. R14 Two participants commented that although it did not concern them, they recognized that data protection was a common anxiety. In con-

| Statement of principal findings
This is the first study in the UK to investigate providing patients with a written summary of their care whilst they are still in hospital. The response from this group of patients was overwhelmingly positive: they reported that it helped them retain more information, provided transparency and improved their confidence in asking questions.
Healthcare professionals reported that the summaries improved their explanations to patients by prompting them to reduce medical jargon and helped to solidify their treatment plan. Concerns about confidentiality breaches were voiced by clinicians but, unexpectedly, these were not shared by patients and no reports of breaches of confidentiality occurred. According to healthcare professionals, the most significant barrier to implementing these summaries is the time taken to write them; they are uncertain whether the benefits described are sufficient to outweigh the extra time burden in the long term.

| Strengths and weaknesses of the study
Most patient and clinician participants who were approached agreed to be interviewed and this included a wide variety of clinicians working on two wards. Nurses were included due to their role as educators and facilitators of patient-clinician communication. 18 Differences between nurses' and doctors' views have been highlighted where present, but these differences were few. The interviews were carried out during the patients' acute admission, so their recollection of events was still strong. Interviews were double coded, providing robust analysis.
The study was limited to one teaching hospital which uses an electronic medical record, so the findings may not be transferable to, or technologically feasible, elsewhere. Our participant responses may not be generalizable to other groups as data on participant characteristics was not collected, however information regarding the index of multiple deprivation (IMD) of the area was included to give an indication of the characteristics of the population involved. While no anxiety was reported by patient participants, clinicians retained discretion not to give the summary where they thought distress would be likely.
No quantitative data was collected, for example, on time taken to write or distribute the summary or the number of questions asked by patients. Future versions of the summary could include adjustments for easy readability, for example, large font. We were unable to explore the experiences of relatives or patients who lacked capacity or did not speak English; future studies should ensure these groups are included. Interviewers were all clinically trained, which may have biased the research. Interviews were not transcribed verbatim, although having two researchers present ensured that one could ask the questions and the other recorded the interview in notes, with some sections recorded verbatim. Coding differences were few and were resolved via discussion; nevertheless, it should be acknowledged that discussions can be limited by power differentials and a bias towards efficient consensus.

| Tailored written information: Knowledge and understanding without anxiety
Our study suggests that patients value a written summary as a memory aid and to help those close to the patient stay abreast of problems. In a study of patients randomized to receive verbal only or verbal and written discharge information in an emergency department, 90% of the group who received written information knew their diagnosis, compared to 73% of the control group. 19 Clinicians expressed concern that providing written information might provoke patient anxiety: in a similar study to ours, conducted in the United States, only 16% of patients reported being worried or confused by the daily notes they were given. 11 In a portal study only 8% of patients using the portals were confused or anxious about the information they saw. 5 In our study, where pertinent information was selected and presented in an accessible manner, no-one reported anxiety, although there were reported concerns about understanding the content; future work should examine understanding of written summaries. The doctors in our study reported that they observed a change in their verbal communication during the study: they used less jargon and thought more about the structure of how they were explaining things, to be consistent with the written information they were about to provide.

| Patient empowerment
Patients reported feeling more confident in asking questions, and appreciated the opportunity to spot and correct mistakes; this is consistent with other research conducted on written records 11,20 and electronic portals. 12,21,22 Although the summary note provided patients with a degree of additional agency, we recognize that there is still an element of 'curating' what information is given, which is more paternalistic than providing patients with their entire notes. The patients interviewed in our study seemed comfortable with this, trusting the doctor to tell them what they needed to know. Providing a summary of records rather than the entire medical record allows doctors to take on some of the emotional work of worry and uncertainty, 23 while still equipping patients with enough information to ask questions and challenge them. It is not clear, however, whether the patients in our study were aware to what extent the information was being selected on their behalf, which reflects a paradox. Patients have the prerogative to choose not to know things (and/or to prefer paternalistic behaviour from their clinicians), but, if they choose to see a summary record rather than their complete medical record, they will never know what they do not know-and so their choice can never be fully informed.
In primary care the effect of sharing medical notes with patients varies: older, less educated, non-white or non-English speaking patients reported the most benefit in a U.S. study, 6,24 perhaps because they could seek help in reading and understanding their notes. 25 In our study we did not evaluate patient demographics as the sample size was too small for meaningful comparisons; any future research must include this analysis.

| Practicality of using summary records
As with other studies, 12,26 doctors noted the increased time taken to write and explain the summary note. In previous studies, this was considered a worthwhile investment because of the perceived improved patient understanding. However, in these studies the unedited medical record was distributed, whereas in our study a separate summary note was written, resulting in 'double documenting'. 27 Without the additional resource of two medical students, the summary note stopped being distributed at the end of the study, suggesting that, in practical terms, the work needed was too much to undertake for the (at the time, intangible) patient benefit observed. The feasibility of maintaining confidentiality with additional pieces of paper being given to patients was also explored. In our study no patients were concerned about this and there were no reported loss of notes or breach of confidentiality-this is consistent with other studies. 11,28 Concerns about breach of confidentiality should therefore not be seen as a significant barrier to information sharing in hospital.

| Unanswered questions and future research
This study demonstrated the impact on patient-clinician communication as a result of sharing a PTWR summary; however, barriers to implementation were identified, in particular the need for extra human resource. The intervention could be further iterated and optimized via collaborative design methodologies 30 with a range of users. A further mixed-methods multi-centre study would be needed to quantify changes in patient knowledge and understanding of their diagnosis and treatment plan, what they remember of what was said at the consultation, satisfaction and engagement in care (as indicated by confidence in asking questions), including self-care behaviours. The effect on different demographics of patients must be included in any future analysis. While health economic analysis would be secondary to investigating whether the note changed patient knowledge or experience, its enquiry would help ensure that the implementation of such an intervention was properly resourced. Finally, patients could be given the opportunity to see both their full records and a summary, in order to be fully informed about what is contained in the medical record and what doctors are selecting to put in the summary, before being asked which method of information sharing they would prefer.
By putting patients at the centre of future evaluations, and looking for unintended as well as intended effects, we can develop a new model of information sharing which benefits patients and clinicians.