Towards patient‐centred communication in the management of older patients' medications across transitions of care: A focused ethnographic study

Abstract Background Communication about managing medications during transitions of care can be a challenging process for older patients since they often have complex medication regimens. Previous studies highlighted that links between communication breakdowns and medication incidents in older patients occur mainly at discharge or in the post‐discharge period. Little attention has been paid to exploring communication strategies facilitating patient‐centred medication communication at transitions of care from a discourse‐analytic perspective. Objectives To explore, through a discursive lens, strategies that enable patient‐centred medication communication at transitions of care. Design A focused ethnographic study was employed for this study. The study was reported according to the COREQ checklist. Methods Interviews, observations and focus groups were analysed utilising Critical Discourse Analysis and the Medication Communication Model following thematic analysis. Data collection was undertaken in eight wards across two metropolitan hospitals in Australia. Results Patient preferences and beliefs about medications were identified as important characteristics of patient‐centred communication. Strategies included empathetic talk prioritising patients' medication needs and preferences for medications; informative talk clarifying patients' concerns; and encouraging talk for enhancing shared decision‐making with older patients. Challenges relating to the use of these strategies included patients' hearing, speech or cognitive impairments, language barriers and absence of interpreters or family members during care transitions. Relevance to clinical practice To enhance medication communication, nurses, doctors and pharmacists should incorporate older patients' preferences, previous experiences and beliefs, and consider the challenges faced by patients across transitions. Strategies encouraging patients' contribution to decision‐making processes are crucial to patient‐centeredness in medication communication. Nurses need to engage in informative talk more frequently when administering the medications to ensure older patients' understanding of medications prescribed or altered in hospital settings.


| INTRODUC TI ON
Transitions of care involve older patients' movements between different settings and they also comprise consultations with various health professionals (World Health Organization, 2019).  (Ellins et al., 2012;Jeffs et al., 2017).
Patient-centred communication, where health professionals engage with patients and their families and provide medication information that they can understand, can help to facilitate patient safety (World Health Organization, 2019). This process comprises health professionals responding to patients' informational needs, preferences and concerns, being sensitive to patients' emotions and beliefs, providing empathy and support, and enabling patients' self-management and involvement in decision-making (De Haes & Bensing, 2009;Epstein & Street, 2007;Ganz et al., 2014;Hashim, 2017). Barriers to patient-centred communication with older patients include cognitive barriers (Grealish et al., 2019), language barriers (Nkrumah & Abekah-Nkrumah, 2019), memory deficits (Anna & Simon, 2018), and ageist or task-oriented communication styles (Ayalon & Tesch-Römer, 2018;Belcher et al., 2006).
Previous studies examining patient-centred communication have included interviews with patients and observations of clinical practice where older patients participated across transitions of care (Feder et al., 2019;Lenaghan, 2019). Studies showed that older patients were an important source of information about medication changes across settings; however, they often had limited opportu- have involved focusing on a single transition point such as discharge (Allen et al., 2018;Rognan et al., 2021;Tobiano et al., 2021), or on interactions with a single health professional group such as patientpharmacist interactions (Braaf et al., 2015) or patient-doctor interactions (Kripalani et al., 2007). Additionally, little attention has been placed on exploring discourses used by health professionals and older patients during medication interactions.
The aim of this study was to explore, through a discursive lens, strategies that enable patient-centred medication communication at transitions of care.

| Research design
A focused ethnographic study was undertaken, with semi-structured interviews, observations and reflexive focus groups. Focused ethnography has evolved as a pragmatic form of conventional ethnography, which enables researchers to investigate a specific social phenomenon in a particular context for an identifiable period of time, and with a pre-established goal in mind (Knoblauch, 2005;Mayan, 2009;Roper & Shapira, 2000). Unlike conventional ethnography, the researcher who uses a focused ethnographic approach examines a distinct problem and concentrates on particular research questions that are formulated before going into the field (Christine et al., 2009;Higginbottom et al., 2013). In focused ethnography, the researcher is familiar with the research context and can conduct short-term field visits supported by audio-visual technologies to collect large amounts of data (Higginbottom et al., 2013;Simonds et al., 2012). The COnsolidated criteria for REporting Qualitative studies (COREQ) checklist for qualitative research has been followed in reporting this study, see Appendix S1 (Tong et al., 2007).

| Study setting
This study was conducted in two metropolitan hospitals comprising an acute tertiary hospital (Site 1) and a geriatric rehabilitation facility (Site 2), in Melbourne, Australia. Data were collected in two specialty medical wards and one general medical ward at the first site and two rehabilitation wards and three aged care wards at the second site.

| Participants and sampling procedure
Patients aged 65 years or older admitted to the study wards were eligible for inclusion; those with severe dementia were excluded.
Purposive sampling was undertaken to ensure adequate representation of older patients from different ages including youngest-old (65-74), middle-old (74-84), and oldest-old (over 85 years). Older patients from non-English speaking backgrounds were included in observations if they were able to verbally consent in the presence of family members or interpreters who helped them with translation.
Almost 21% of older patients who participated in the observations were of non-English speaking backgrounds. The health professionals included were nurses, pharmacists, and qualified doctors who worked in study wards at least 1 day per week. Casually employed nurses were excluded. Family members of patients who were involved in the care of the older patients, and who spoke and understood English were considered for inclusion such as older patients' children, siblings, spouses or friends. After completion of interviews and observations, focus groups were conducted with different older patients and family members to those involved in interviews and observations. Two field researchers undertook the data collection. Patient lists were reviewed by researchers for suitable patients to be approached. These lists included information about patients' ages, gender, the admission date and reasons for hospital admission.
Researchers liaised with bedside nurses and nurse unit managers to determine older patients' eligibility. Older patients were recruited after researchers provided verbal and written information about the study. Sampling of health professionals occurred purposively, based on their professional disciplines, and the length and level of professional expertise in different settings. Approximately more than half of participants who were approached consented to take part in the study.

| Data collection
Data collection consisted of semi-structured, face-to-face interviews and participant observations at both hospital sites from April 2018 to October 2019. While semi-structured interviews were conducted with all participant groups, researchers were particularly interested in older patients' perspectives of communicating about medications across transitions of care (Table 1). An approximate length of an interview was 20 min.
Observations were conducted during key medication-related activities, such as medication administration, ward rounds, handovers, admission and discharge sessions ( Table 1). The observational data were captured using audio recordings of researchers' observation and field notes. The researchers also took photos and documented older patients' medications from the patient medical records. After completion of interviews and observations, focus groups were undertaken at both hospital sites with different older patients and family members to those who participated in the initial interviews and observations. Focus groups were conducted to obtain patients' and family members' reflections on the key issues identified during interviews and observations. The research team determined the key issues following preliminary thematic analysis of interviews and observational data and formulated focus group schedules based on these issues. Focus groups were conducted with different patients and family members to the participants who were recruited for interviews or observations. Audio recordings were collected on all data and transcribed verbatim, and summary field notes were made after each observation.
The two researchers who collected data maintained detailed field notes and began coding during data collection. Data were collected in one ward setting at a time in order to comprehensively explore the specific contextual characteristics of a specific ward setting. Field notes and codes were regularly examined during data collection. Through conversations with the whole team and reflections on field notes and codes, data collection ceased within a particular ward setting if repeated patterns of practices, experiences and perceptions were found. This occurrence of repeated patterns of practices, experiences and perceptions was viewed as the point where data saturation had been reached. The two researchers then moved to the next ward setting for data collection. Two field researchers, the study's first author and a research fellow, from pharmacy and nursing backgrounds collected data respectively. Both researchers were female and received training in qualitative research methodology at PhD and Masters level. The researchers introduced themselves as scholars to older patients, with an interest in hearing their experiences of communicating about managing medications across transition points of care. The researchers did not have a prior relationship with any participants in this study.

| Data analysis
Thematic analysis of transcriptions was initially undertaken (Braun & Clarke, 2006). It was an iterative and content-driven approach, which required reading and re-reading transcripts, listening to audio files and exploring key ideas and themes by the two field researchers. Initially, verbatim transcriptions of interviews, focus groups and observations were imported into NVivo 12 (QSR Melbourne) for data management. Then, two researchers coded each transcripts independently by reading through all transcripts line-by-line. Thematic analysis helped the research team to make sense the data by exploring the initial codes. The  (Fairclough, 1992) • Who was talking or silent in medication communication? How was body language used? How were language devices such as hedging (words used that weaken the force of the statement), modality (words that express the degree of certainty, possibility, necessity or willingness), turn-taking (patterns of speech in which speakers talk one at a time in alternating turns) used by participants? research team met fortnightly to discuss findings and identify commonalities across data. These commonalities were collated by the research team into potential themes. Once potential themes were confirmed, the researchers worked through the themes to identify the different analytical elements relating to the three dimensions of Critical Discourse Analysis (CDA) and the Medication Communication Model (Fairclough, 1992;Liu et al., 2011;Manias, 2010).

TA B L E 1 Interview, observation and reflexive focus group schedules
The three dimensions of CDA comprising the discursive practice level, text level, and social practice level were mapped against three levels of the Medication Communication Model (Manias, 2010), which consisted of antecedents, attributes and consequences respectively. At the discursive practice level, we investigated how the text was produced, communicated, and consumed by individuals. At the text level, we identified the content and structure of the text, and explored language devices, which helped to understand how health professionals including nurses, pharmacists and doctors communicated in response to older patients' preferences, needs and beliefs. At the social practice level, we identified implications of social relations on discursive practices (Table 2). Thematic analysis and critical discourse analysis were congruent with each other, and their combined use facilitated a complimentary approach to data analysis. Initially, thematic analysis enabled us to inductively identify patterns in the data, and to make sense of the commonalities and differences obtained in episodes of social interactions between individuals. Subsequently, critical discourse analysis was viewed as a means of examining the themes in diverse ways to determine influences affecting communication, the content and nature of communication, and sociocultural and health consequences, such as patient participation in decision-making processes and power disparities.
The research team created a codebook and guide questions that were formulated to facilitate the data analysis process (

| Ethics approval
This study was approved by the health service ethics committee of the two hospital sites, HREC 212/17 and the university committee, DUHREC 2018-067. Researchers obtained informed written consent from older patients, family members and health professionals.
To maintain the confidentiality, pseudonyms were used in data excerpts.

| RE SULTS
Interviews were undertaken with 50 older patients and 29 health professionals including doctors, pharmacists, and nurses. In total, 203 h of observations were conducted with 29 health professionals and 111 older patients. Focus groups were conducted with 20 patients and 13 family members. Results were analysed from patient interviews, observations of the interactions occurred between health professionals, older patients and family members if they were present as well as focus groups with older patients and family members. Figure 1 shows the key findings. The patient perceived that using a pharmacist-prepared dose administration aid would compromise his autonomy in managing medications following discharge. His statement: 'I feel I'm in control', indicated that he felt safer not using the aid. The patient's attempt to be actively involved demonstrated the tension between the patient discourse of self-control and the pharmacy discourse of preventing harm.

| Patient characteristics affecting individualised medication communication
Observations revealed that older patients were admitted to acute care settings in order to undergo a specific procedure or to stabilise their acute or chronic medical conditions. Therefore, the time that they spent in these settings was usually limited compared to those patients who were admitted to geriatric rehabilitation settings to receive more specialised and long-term treatment to manage their complex and multiple conditions before discharge home or to residential aged care. Therefore, the time that pharmacists spent with older patients in geriatric rehabilitation settings was relatively longer than with those who were admitted to acute care settings. Within geriatric rehabilitation settings, pharmacists were able to dedicate more time with these patients to discuss about their medications.
Additionally, pharmacists' education in acute care settings tended to be more rushed due to the relatively high flow of patients who were just admitted or those who needed to be discharged.
The patient's resistance to taking the vitamin B12 injection was associated with her belief that an injection could cause an allergic reaction. Since the patient had previously experienced an allergic reaction from a penicillin injection, she was reluctant to have further injections. Nursing and medical staff reportedly did not perceive that conveying information to the patient about the vitamin B12 injection was important.

| Establishing patient-centeredness in medication communication
The strategies that health professionals used to establish patient-

*Excerpt Code: Pharm5_Obs_Subacute2_
Interaction15_ Pt110 The insulin dosages that the medical team prescribed were slightly lower than those the patient had previously taken at home.

| Informative talk provided by health professionals
Informative talk occurred when health professionals provided older patients with tailored education about medications in a way that they could understand. Older patients reported needing regular discussions with health professionals about medications, particularly when they were given a new medication after hospital transfer. Some patients felt dissatisfied with the amount of information given and they reported that they did not receive the information they required. These patients involved those who managed medications independently at home and those who felt responsible in keeping track of the medication changes that occurred across transitions. These patients prompted health professionals to provide more information about medications. Some who were admitted to the acute settings with complex health conditions and complained about memory issues indicated that they preferred not to have an informative talk with health professionals at hospital admission since they were too sick to retain information. However, they wanted information once their health started to improve, and they became more capable of discussing their Observations revealed that nurses' use of informative talk rarely happened during medication administration. Nurses prioritised their efforts in preparing and administering medications rather than providing informative talk during these medication activities. Our observations explored that nurses tended to engage with informative talk when they administered particular medications including laxatives, painkillers or inhalers. Because if that's really compacting the tummy it can make your pain worse. Now and I'll also have a look, is there any more pain medication I can give you.

Patient:
And the lady (refers to the pharmacist) that came in saying, she is the pharmacist, she was talking about some lozenges for nicotine treatment.
Registered Nurse: Ok, yeah. (Pause) Do you want a break from the oxygen? The nasal prongs?

*Excerpt Code: RN1_Obs_WardMed1_
Interaction10_Pt12 The nurse talked on behalf of the doctors to provide the rationale behind why they prescribed Coloxyl ® and Senna (docusate sodiumlaxative) upon the patient's admission to hospital and emphasised its importance as part of pain management. While informing the patient, the nurse frequently used the plural personal pronoun (i.e. 'we'), which indicated adoption of a partnership discourse to convince the patient to take the tablets. However, the nurse dominated the talk by having interactional control over the medication conversation and no effort was made to include the patient. In this interaction, there were conflicting medication interests between the patient and the nurse since the nurse was trying to provide information about laxatives but the patient was interested in obtaining further information about the lozenges for nicotine treatment.

| Encouraging talk to enhance shared decision-making
Encouraging talk referred to using inclusive communication strategies with a focus on motivating older patients to be involved in medication decisions. Health professionals engaged in encouraging talk particularly when they wanted older patients to take responsibility in managing their own medications, especially following discharge.
Health professionals' encouraging talk occurred with older patients whose medications needed to be altered after discharge, with patients whom pharmacists believed would require dose administration aids to support the taking of their medications in the future, or with patients who were hesitant in taking particular medications.
Observations revealed that encouraging talk rarely happened with older patients who were slightly confused, or not fluent in English or who had hearing impairments. With these patients, health professionals attempted to involve family members if they were present at the bedside.
Doctors encouraged older patients to be involved in decisions about medications by seeking their preferences in situations where doctors made changes to patients' regular medications or when they decided to prescribe a new medication for acute symptoms such as pain or constipation: Registrar: We can start some Fibogel ® or Metamucil ® .
(to patient) Have you ever used Metamucil ® ?
Patient: Yeah, I've got it at home.

Registrar: Do you find it works?
Patient: I can't recall.
Registrar: Will we try it?

Patient: (responds affirmatively)
Registrar: You've got to have a lot of water with it.

Patient: Is it orange?
Registrar:…might be a different brand but same sort of thing… *Excerpt code: Med6_Obs_ Geri_Rehab4_

Interaction_7
In this excerpt, the patient was a 78-year-old man with constipation. Doctors consulted with the patient about the plan for laxatives.
The registrar used the questioning strategies to encourage the patient to provide more information about his experience of using laxatives at home. Although the patient did not recall his previous experience with Metamucil ® , the registrar used a partnership discourse (Will we try it?) to be more inclusive for the patient to participate in decisionmaking. Although the registrar engaged in encouraging talk, it was not clear that this attempt led to shared decision-making between the doctor and the patient, because the patient's contributions to the conversation remained very minimal. The doctor's use of closed-ended questions only required the patient to approve or disapprove his suggestions rather than facilitating further input from the patient.
Nurses engaged with encouraging talk when older patients sought to avoid taking particular medications or when nurses wanted older patients to take responsibility to express their medication concerns or when nurses asked older patients to direct  English . Competing interests between patients and health professionals, health professionals' multitasking activities or older patients' preference to take a passive approach in med- nurses were observed to prioritise the discourse of task completion over the discourse of eliciting preferences before discharge, which prevented them from taking a patient-centred approach.

| DISCUSS ION
Nurses also perceived that communication about medication with patients after admission or before discharge was the pharmacists' role rather than their role (Bolster & Manias, 2010

| Limitations
A limitation of this study is the potential for the observer effect to occur in participating health professionals during their communication with patients. However, researchers spent extensive periods with participants, which helped to gain rapport and familiarity.
Researchers also emphasised their intention to understand complexities with medication communication in older patients rather than judging practice.

| CON CLUS ION
Our findings elaborated patient preferences, experiences and medication beliefs as characteristics that need to be considered by health professionals when providing individualised medication communication to older patients in the context of care transitions. Shared decision-making was more likely to happen when health professionals adopted encouraging talk where they used a partnership discourse and inclusive statements during medication communication. Engagement with empathetic talk during transitions of care was helpful for health professionals to reduce unequal power relations between themselves and older patients, which, in turn, enabled older patients to communicate more confidently about their medication preferences.

| RELE VAN CE TO CLINI C AL PR AC TI CE
While communication skills are taught in university programmes, health professional students are often not taught about the communication skills relating to helping patients and families to manage medications in clinical practice across transitions of care, as demonstrated by the missed opportunities identified in the study (Manias & Bullock, 2002). Health professionals should identify older patients' preferences for both managing medications and including themselves in medication communications at different time points during their movement from admission to discharge. In some cases where the patient turnover is high such as in acute care settings, health professionals need to be aware of that they do not prioritise the discourse of organisational efficiency over the discourse of eliciting older patients' medication preferences. Health professionals need to organise education sessions with family members or interpreters for older patients with language barriers or cognitive issues. Organisational commitment is required to explore older patients' beliefs regarding medications because their beliefs can determine whether patients would take and manage their medications correctly during the post-discharge period. It was observed that the healthcare environment involves many organisational ten-

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest.

AUTH O R CO NTR I B UTI O N S
GO and EM made substantial contribution to designing and drafting of the manuscript. GO, EM and were responsible for data collection/ analysis and data interpretation. EM, TB, CJ, CH, RW and made critical revisions to the paper for important intellectual content. EM, TB, CJ, CH and RW obtained funding. All authors have agreed on the final version of the review.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.