Finding the Person in Electronic Health Records. A Mixed-Methods Analysis of Person-Centered Content and Language

ABSTRACT

The co-creation and sharing of documentation of person-centered care is important, but challenging in clinical practice. Online access to health records is considered essential to increase patient participation and empowerment in person-centered care provision. The aims of this study were (1) to identify the extent of person-centered content in medical inpatient records and discharge letters; (2) to describe the characteristics of the language and rhetoric used in discharge letters. This was a concurrent, mixed-methods study involving a deductive, quantitative analysis of person-centered content in 69 Patient Accessible Electronic Health Records from patients with pituitary tumors, and an iterative, qualitative language analysis of a nested sample of 57 discharge letters. Both the content and language of inpatient records were mostly patient-centered. Records were concerned with the documentation of symptoms and medical and care interventions. There was little person-centered documentation of patients’ preferences, wishes and needs, and shared decision-making. In the discharge letters, written for the patients, some physicians had attempted to write in a person-centered way, using plain language and a narrative. However, most wrote in a style that was reflective of their discourse community, using abbreviations and medical terms. Established norms for documentation in healthcare are a barrier to person-centered documentation. Patients’ needs and preferences about documentation should be explored to increase understanding of how person-centered documentation can be achieved in clinical practice.

Introduction

Person-centered care (PCC) is grounded in the ethics of personhood. It aims to humanize healthcare provision through active patient involvement and partnership between healthcare professionals and patients (Ekman et al., 2011; McCormack & McCance, 2006, 2017).

Ekman et al. (2011) described three elements of PCC: (1) eliciting the patient’s narrative of their personal illness experience, and their wishes and needs, (2) forming an equal partnership, and (3) co-creating and sharing health documentation. In long-term and complex health conditions in particular, including patients’ illness experience and life situation in care planning can help them to self-manage their illness (Durand et al., 2014). Shared decision-making can also reduce health inequalities (Durand et al., 2014). Patients have individual and changing preferences about their desired level of involvement in shared decision-making. However, they generally value information and the humanity of person-centered relationships (Wolf et al., 2017; Zizzo et al., 2017).

Sharing written information through Patient Accessible Electronic Health Records (PAEHRs) is important in increasing patient participation and improving care quality (De Lusignan et al., 2014; Wiljer et al., 2008). PAEHRs enable patients to read and sometimes edit or comment on their records (De Lusignan et al., 2014). They are used in many countries including Australia (Pearce & Bainbridge, 2014), Canada (Wiljer et al., 2008) and Denmark (Nørgaard, 2013).

In Sweden, all patients have had online access to a PAEHR via a national patient portal since 2017. The Swedish PAEHR includes clinical information such as diagnoses, examinations, care and treatment plans, medication, referrals, letters to specialists, and discharge letters about inpatient and outpatient episodes (1177Vårdguiden, 2020).

Patients appreciate having access to their health records (Gerard et al., 2017; Zanaboni et al., 2020). However, sharing PAEHRs does not in itself promote patient participation in care planning or decision-making (Murphy et al., 2016). Person-centered documentation is a significant challenge (Britten et al., 2017). This manuscript focuses on two specific issues related to this. First, healthcare professionals are uncertain what content should be documented in PAEHRs to show the patient as a person, the holistic aspects of care provision, and how patients participated in decision-making (Dellenborg et al., 2019; Jakobsson et al., 2019). A person-centered relationship in clinical practice is not necessarily documented because health records templates often prompt symptom reporting rather than biographical or personal details (Broderick & Coffey, 2013).

The second issue is how to document. Healthcare professionals often write in a technical-scientific style (Hyde et al., 2005). Patients may struggle to understand medical terms or phrases (Zanaboni et al., 2020), instead of plain language (Walsh et al., 2017).

The concept of discourse communities (Paré, 2017; Swales, 1990, 2016) helps to explain these issues. The term discourse relates to written and spoken communication. Discourse can confirm and change social relationships and behaviors (Burr, 2015; Lemke, 1995). Specific discourses occur in so-called discourse communities, which are defined by a shared purpose or interest, and the development of a community-specific vocabulary, communication styles, norms, and topics. Registered nurses or physicians share writing styles, which include acronyms, abbreviations, and technical, task-focused language (Anspach, 1988; Broderick & Coffey, 2013; Paré, 2017; Swales, 1990, 2016). This language is suitable for effective and task-oriented inter- and intra-professional communication within the discourse community, but arguably not for sharing information with patients. The majority of patients are not part of the medical or nursing discourse community. Less well-educated patients may find it particularly difficult to understand medical notes (Gerard et al., 2017; Mishra et al., 2019).

Increasing patient participation in healthcare requires the discourse community of health professionals to include PCC content in medical records and to develop an appropriate language to communicate with patients. However, knowledge about how to achieve this is scarce. This study therefore aimed to analyze the PCC content of PAEHRs and examine the language used in discharge letters. Discharge letters were used because, unlike other medical documentation written for professionals, these letters are written by physicians for patients.

Method

Research design

This was a concurrent, mixed-methods study (Creswell & Plano Clark, 2010) including a quantitative and a nested qualitative analysis (Onwuegbuzie & Collins, 2007). We analyzed two aspects of person-centered care documentation: the person-centered content of PAEHRs and physicians’ use of language in a nested sample (Onwuegbuzie & Collins, 2007) of discharge letters to patients. The qualitative and quantitative analyses were conducted independently and synthesized at the results level (Figure 1).

Figure 1. Study design.

Setting and context

This study is part of a larger research project at Sahlgrenska University hospital in Sweden (Jakobsson et al., 2020). The PAEHR excerpts were from medical records of inpatient episodes, including a specific patient-focused section, the Care and Medication Letter (also known as a discharge letter). This is written by the treating physician to the patient, and is nested within the medical record. The letter summarizes the inpatient care and treatment, and the follow-up arrangements, covering the reason for admission, current drug prescription, care and medication, further planning, responsible physician, and contact. The discharging physicians are free to complete or omit any of this information. Patients receive a hard copy of the letter on discharge from the ward. The PAEHRs were produced in a ‘natural context’, a hospital ward where some staff had received PCC training, but PCC had not been implemented systematically. Only nurses’ and physicians’ entries were included in the analysis.

Sample and sampling

The sample included 71 medical records of postoperative inpatient care episodes of consecutive patients with pituitary tumors from an 18-month period (June 2016–November 2017). We excluded two records that were not available to the research team. The final sample contained 69 medical records, with a nested sample of 57 discharge letters (Figure 1).

Ethical considerations

The Regional Ethical Review Board in Gothenburg, Sweden, approved the study (registration number 387–15). All data were treated confidentially; all personally identifiable data were removed from the medical records.

Data analysis

We conducted equally-weighted concurrent quantitative and qualitative data analyses. The quantitative analysis examined the extent and types of PCC entries in medical records and discharge letters. The qualitative analysis examined the characteristics of the language in the discharge letters.

The PAEHRs were extracted from the hospital’s electronic health record system. Staff and patient details were irreversibly de-identified using Adobe software. The discharge letters were transcribed and imported into NVivo 12 Pro (QSR International, released 2018) for further management and analysis.

Quantitative analysis of content

The quantitative analysis of medical records used a template of three domains and ten subdomains (Jakobsson et al., 2019). Domain I (Knowledge of the patient as a person) included subdomains (a) the patient as a person, (b) preferences, (c) resources, (d) physical symptoms, and (e) psychosocial situation. Domain II (Provision of holistic care) covered (a) patient’s resources, and (b) preferences/wishes. Domain III (Partnership in care and planning) included (a) information provided, (b) participation in discussion, and (c) participation in decision-making (Jakobsson et al., 2019).

The records were rated on the extent to which each subdomain was present on a five-level Likert-type scale ranging from “not at all” to “a great extent”. MC was the main coder, and BH reviewed the data. They critically discussed the coding and any discrepancies during seven meetings at regular intervals. Overall, the interrater agreement was high, with Cohen’s Kappa ranging from 0.80 to 0.98 for all categories.

Qualitative analysis of language

Discharge letters were analyzed linguistically in a qualitative content analysis (Schreier, 2012) with initial categories derived from the literature. We explored the language use based on Anspach’s (1988) four categories of technical-instrumental writing. These categories are (1) de-personalization, (2) the use of agentless passives, (3) the description of technology as an agent, and (4) account markers. Account markers such as ‘the patient states that … ’ suggest doubt about the veracity of a patient’s narrative. We also used a set of mirror categories covering a more personal writing style. The more person-centered categories were personalization, active agents, technology as a supporting tool, and account markers that indicate acceptance of patients’ narrative. The grammatical structure was examined using the structured and unstructured expository writing characteristics set out by Arntfield and Johnston (2016). They defined expository writing as a focused summary or explanation of an event or theme, presented in a reasoned, informed, and factual fashion. Structured writing includes full sentences containing a subject, a verb and objects, and is conversational. Unstructured writing features point form or scratch notes (Arntfield & Johnston, 2016).

The qualitative analysis was an iterative process over three cycles. BH initially coded the data deductively by applying an initial coding frame. SJ and BH reviewed and discussed the coding during two meetings. The first meeting was organized after the second cycle of coding, and the second finalized the coding frame. The smallest unit of analysis was single words such as ‘you’ or ‘we’ combined with a verb to indicate personalization or de-personalization, and entire sentences (subject, verb, and object) to analyze the overall rhetoric.

Results

Demographic data

Our sample included 69 medical records of 30 women and 39 men. The patients’ mean age at recruitment was 59 years (range 28–85 years). There were 65 patients (92%) with a benign pituitary adenoma, two with craniopharyngioma tumors, one with a benign brain tumor, and one with a malignant tumor of the pituitary gland. The length of inpatient stay was 1–26 days (mean 7 days).

Medical records and discharge letters: Person-centered content

There were some person-centered entries in records. Some domains were more frequently reported than others (Table 1).

Table 1. Degree of person-centered content in electronic health records

		Not at all N (%)	Somewhat N (%)	Quite a lot N (%)	Well-documented N (%)	To a great extent N (%)	Mean rating
Domain	Subdomain
Knowledge of the patient as a person	a) Person	3 (4)	34 (50)	24 (35)	5 (7)	3 (4)	1.6
	b) Preferences	29 (42)	25 (36)	13 (19)	2 (3)	0	0.8
	c) Resources	0	0	13 (19)	45 (65)	11 (16)	2.9
	d) Physical	0	1 (1)	9 (13)	43 (62)	16 (23)	3.1
	e) Psychosocial	7 (10)	29 (42)	27 (39)	4 (6)	2 (3)	1.5
Provision of holistic care	a) Patient’s resources	56 (81)	10 (15)	3 (4)	0	0	0.2
	b) Patient’s preferences/wishes	51 (74)	14 (20)	2 (3)	1 (1)	1 (1)	0.4
Partnership in care planning and provision	a) Information provided	4 (6)	4 (6)	14 (20)	33 (48)	14 (20)	2.7
	b) Participation in discussion	48 (70)	16 (23)	2 (3)	2 (3)	1 (1)	0.4
	c) Participation in decision-making	51 (74)	13 (19)	3 (4)	2 (3)	0	0.4

Notes: n = 69

The subdomains patient’s resources (Ic), physical problems (Id), and provision of information (IIIa) were mostly ‘well documented’ or reported ‘to a great extent’. All other categories were reported to a limited degree.

Patients’ psycho-social status (Ie) was documented in some cases. Examples included “[Patient] feeling disappointed as the situation has not improved after surgery” (ID 65), “His wife has been with him [patient] the whole time, which is nice as he does not want to be alone because he suffers from panic attacks” (ID 62), and “The patient is suspicious in his social relationships” (ID 43).

Some entries were concise, but gave an impression of the patient as a person. Examples included “the patient lives in an apartment with her partner and her two children. Works full time with research, physically active with running, and cycling if she needs to get somewhere” (ID 58) (Ia, Ib, Id, Ie).

The subdomains holistic care provision based on patient resources (IIa) and preferences (IIb) were mostly not documented. Patient participation in discussion (IIIb) and decision-making (IIIc) were either not or only somewhat documented.

PCC provision covers diagnosis, analysis, and decision-making. Some entries concisely captured this process and the involvement of the patient, for example: “PM: [patient] does not want lactulose, [I] gave laxeriva [laxative] instead” (ID 61). This short entry covers the patient’s wish (IIb), the discussion with staff (IIIb), and the decision (IIIc). In most other records, it was unclear whether patients’ preferences were considered.

In a few medical records, the documentation was more person-centered. In these cases, the patients or their relatives had asked questions or wanted to discuss care and treatment, taking active charge of the situation. One patient who was terminally ill with cancer clearly communicated his wishes throughout his stay on the ward. Another patient asked to speak to physicians: “AM: Pat[ient] is well but has several questions and thoughts about lab results and hormone substitution. Will speak to physician. Is very happy after speaking with the physician. […] PM: Pat[ient] wondering if he will meet the operating surgeon for a conversation before discharge. Will take this up on ward round. […] AM: Pat[ient] has questions about testosterone levels and reference levels […]” (ID 48).

Discharge letters: Language and rhetoric

All entries were expository. The majority featured a mix of structured and unstructured writing. Ten were entirely unstructured, and one was entirely structured. The letters varied in length from 65 to 457 words (mean: 191.3). Table 2 shows the final coding framework.

Table 2. The final qualitative coding framework: language, rhetoric, number of codes and patient records

Category	Code Counts in discharge letters (N)	Sub-Code Discharge letters (N)	Examples
Rhetoric	Expository structured (46)		‘You will be called for a return visit to ward xx in 3 months to check your pituitary hormones. You will receive a call for a magnet camera examination which will take place in 6 months.’ (ID 36)
	Expository unstructured (56)		‘-Blood pressure control at GP surgery -visit ophthalmologist Monday next week -decision on radiation treatment at the oncologist’s -return visit in 2 months.’ (ID 32)
Biomedical language	De-personalization (43)		‘Reason for admission: Postoperative care after reoperation’ (ID 100)
		Abbreviations, medical terms (40)	‘Transphenoidal total extirpation of microadenoma in pituitary gland.’ (ID 79) ‘…discrete visual field loss upwards temporal.’ (ID70)
	Agentless passives (50)		‘[…] X-ray on 18/11 and one saw a growth of the pituitary gland.’ (ID102)
	Technology as agent (7)		‘The postoperative MRI scan shows…’ (ID 32)
Person-centered language	Personalization (54)		‘You did not feel well at the beginning of your ward stay. You had a headache and nausea.’(ID 100)
		Plain language (30)	‘you have been on […] Minirin tablets because of Diabetes Insipidus, that is a lack of a pituitary hormone called ADH, which causes you to urinate a lot.’ (ID40)
	Active agents (39)		‘You should go to your health center on Monday and Friday next week and have blood samples taken, which they will fax to us. The doctor here will check them and call you if you need to change your Minirin dose.’(ID 82)
		Patient perspective (5)	‘If you urinate a lot at night, we can discuss an increase [Minirin] to 1 tablet again’ (ID87)

De-personalization versus personalization

In most cases, the reason for admission to the ward was stated in a depersonalized manner such as “control after pituitary operation” (ID 33) or “Surgery pituitary tumor” (ID 39). These entries emphasize the surgical procedure rather than saying anything to or about the person who had undergone the operation. Physicians frequently both personalized and objectified patients, for example: “You were operated on the 8/12 for a pituitary adenoma” (ID 57).

Abbreviations and medical terms such as MR (magnetic resonance imaging), ÖNH (Swedish abbreviation for ear, nose and throat), KBT (Swedish abbreviation for cognitive behavioral therapy), and transsphenoidal pituitary surgery were present in most records. However, the use of plain language instead of medical terms and abbreviations helped to personalize the entries. Several plain language entries contextualized, explained, and narrated the situation in simple terms, such as, “You have undergone pituitary surgery because a tumor was detected in the pituitary gland that grew above the pit in the bone [Sella Turcica]. This compressed the optic nerve crossing [chiasma] that lies immediately above the pituitary gland, which has affected the visual field” (ID 70).

Another entry explained monitoring for diabetes insipidus, a common side-effect of operations in patients with craniopharyngioma, who drink and urinate large quantities because of hormonal imbalances. The entry read, “We also measured your fluid balance by writing down how much you drank or peed, etc. [...] and you have passed slightly larger amounts of urine, just as expected after pituitary surgery […]” (ID 62).

Several records described the follow-up arrangements in an unstructured format and using biomedical language, as if directed at colleagues, for example, “The plan: Blood pressure control at VC [primary care clinic], visit to ophthalmologist Monday next week, assessment for radiation therapy by the oncologist, return visit in 2 months” (ID 32). In contrast, other entries featured cohesive, structured, and personalized language to describe the post-discharge arrangements: “Planning: You should go to your health center on Monday and Friday next week and have blood samples taken, [the results of] which they will fax to us. The doctor here wants to check them, and will call you if you need to change the Minirin [medication for diabetes insipidus] dose. On Friday, they can remove the stitches from your leg. In about 1 month, you will be called back to us for blood sampling and a doctor’s visit. In about 3 months, you will have a new MRI of the brain. I have sent a referral to XX, so that you can have the investigation there, but you will get an answer [result] from us. After the summer break, you should be called to [see] our ophthalmologists, who are specialists in visual disturbance. You will be followed up by hormone physicians [endocrinologists] for life.” (ID 82).

Agentless passives versus active agents

The analysis showed a mix of passive and active voices in the discharge letters. Treatment planning or delivery was often described using agentless passives, where the person either initiating or conducting the action was not identified. Examples of agentless passives included: “an electronic prescription was sent” (ID 38), “Testogel dose was changed” (ID 47), and “Visual acuity and field of vision have been examined without any abnormal findings” (ID 61). However, there were also many examples of the active voice. These referred to patient actions or treatment planning or delivery; for example, “You came to the ward after surgery for an unspecific tumor in the pituitary gland. We have followed up hormone tests, measured blood pressure and fluid balance – all to detect any hormone deficiencies after surgery. […] On Monday, you should submit samples at your health center. In 1 month, we will take control samples [to check] for hormone deficiencies. […]” (ID 99). These entries clearly show both patients and physicians as active agents, but “we” is used for ward healthcare staff in general. This implies that an anonymous actor is representing the organization.

Overall, the communication was one-directional, from physician to patient. Patients were the receivers of instructions and information, and their perspective was mostly absent. However, some entries suggested a dialogue or shared decision-making: “We decided together that you should receive testosterone injections instead of gel in the future” (ID 40), or “On 23/1, you also met a counselor to discuss the best way to proceed about your mental health, and you decided that CBT [cognitive behavioral therapy] seems to be the right way for you to go” (ID 62).

Synthesis

The PAEHR content and discharge letter language were mostly patient-centered, and person-centered elements were limited. Patients’ resources, physical problems, and information provision were generally well documented. However, there was little information about their psychosocial status, preferences, needs, and inclusion in decision-making. In the discharge letters, patients were usually simply the receiver of instructions and information. However, the language was often somewhat person-centered, and patients were frequently addressed directly as “you”, which was consistent with the mix of patient and person-centered elements found in the quantitative analysis. The degree of patient- or person-centeredness depended on the individual physician. Some physicians used a narrative style and plain language, and others wrote in a standard, biomedical format.

Discussion

Our study analyzed person-centered content and language in PAEHRs. Few records contained much person-centered content. However, the scarcity of person-centred content does not enable us to draw any conclusions about the person-centeredness of the relationship between patients and practitioners in clinical practice.

This mixed-methods analysis of 69 medical records and 57 discharge letters showed that PAEHRs seldom contained information about patients as persons including their psychosocial situation, wishes, needs, and preferences. We also found little documentation of patient involvement in care-related discussion and decision-making. This suggests that the communication norms of healthcare professionals’ discourse communities may be a barrier to person-centered documentation.

There was little documentation of partnership and participation in care planning and decision-making. This is in line with published research (Broderick & Coffey, 2013; Hyde et al., 2005; Jakobsson et al., 2019; Olsson et al., 2013), but it also highlights a deeper problem: the conflicting goals of PCC and the current bio-medically oriented and patient-centered healthcare system. The bio-medical system aims to enable a functional life (i.e., a focus on symptom management), but PCC focuses on a meaningful life (Håkansson Eklund et al., 2019). This has implications for PAEHRs. Patient-centered documentation is about the patient and usually written by and for healthcare professionals. However, according to PCC, medical records should be produced collaboratively and serve both patients and healthcare professionals (Ekman et al., 2011). The absence of documentation of patient participation suggests that despite the prevalent discourse on PCC, most professionals see PAEHRs in a “traditional”, patient-centered way, as a tool to structure, plan and monitor the patient trajectory (Inera, 2018).

Our analysis of discharge letters showed that the language was mostly technical, with patients being receivers of instructions and information. This informative and expository style of writing is useful and efficient within the discourse community of healthcare professionals. However, patients, particularly those without a university education, may not readily understand their PAEHR (Zanaboni et al., 2020). Patients find it beneficial to take part in person-centered care planning (Wolf et al., 2017) and read their PAEHRs (Mishra et al., 2019; Zanaboni et al., 2020). A study found that patients who experienced PCC, including the documentation, felt more informed and secure (Alharbi et al., 2014). The same study found that patients wanted their personal preferences considered in care planning, provided this was medically appropriate. However, this was a small qualitative study and involved only 16 patients. Further research from the patient perspective on the usefulness of person-centered documentation in PAEHRs is needed (Murphy et al., 2016).

Generally, physicians addressed patients directly as “you”, which personalized the letters. They referred to themselves as “we”. However, many also used depersonalizing language elements and agentless passives. An Australian study also highlighted the lack of plain and personalized language in PAEHRs (Walsh et al., 2017). Language that eliminates actors and uses a passive voice places physicians, their knowledge, and their decisions beyond scrutiny, because they become impersonal entities acting on behalf of the organization (Anspach, 1988). Most physicians mixed personalizing and depersonalizing language. They often addressed patients as “you”, but also used abbreviations and medical terms. The use of technical terms and abbreviations is not surprising, because physicians have a heavy workload, and the habitual use of abbreviations saves time. Patient-centered documentation relies heavily on pattern recognition, and the use of abbreviations also helps other members of the discourse community to recognize situations rapidly (Prior, 2003). However, abbreviations are not consistent with person-centered reporting, because they make it more difficult for patients to read and understand their PAEHRs. Our analysis therefore confirms that, despite the popular discourse about PCC, patients are not usually seen as part of the healthcare community, at least in documentation. Interactions between healthcare professionals and patients may be person-centered, but the documentation is not. Sharing and co-creation of PAEHRs requires patients to be admitted as new members of an established discourse community.

Third, the language analysis showed that it is challenging for practitioners to find a suitable language that matches PCC requirements. The degree of patient- or person-centeredness varied by physician, with some trying to write in plain and accessible language. This shows their general awareness of PCC, and readiness to try to translate their perception of person-centered documentation into clinical practice. Some authors have recommended that assessment and documentation tools are adapted to encourage the inclusion of person-centered content (Broderick & Coffey, 2013). However, we maintain that it is currently impossible to define the required adaptations, because we do not know patients’ needs and wishes about co-creating notes (Murphy et al., 2016). Research that explores person-centered documentation from the patient perspective is urgently needed to enable the development of evidence-based guidance that facilitates successful person-centered documentation.

Strengths and limitations

We used the discourse community concept (Paré, 2017; Swales, 1990, 2016) to frame and explore the challenges associated with person-centered documentation. This concept and its robustness as a social construct may be questioned, but we hold with Swales (2016, p. 10) who argued that “this probably does not matter as long as our focus is on rhetorical principles of organization, on discoursal expectations, on significative linguistic tokens, and on intriguing textual extracts. Such attention on these more surface features provides insight into what at first sight might seem standard, ordinary and predictable”. Our analysis went beyond confirming the well-known lack of PCC in documentation by detecting openness toward practicing PCC, because some physicians tried to write in plain, personalized and accessible language.

The medical records were generated in a “natural context” of a hospital ward, where some staff had received PCC training, but PCC had not been implemented systematically. A potential weakness of our study is therefore that not all staff could be expected to provide PCC. However, we consider this real-life context to be a strength. Some studies on person-centered documentation were conducted during PCC implementation (for example, Jansson et al., 2018). These studies may provide insight into how PCC can function in particular settings, but they do not provide information about the sustainability of the implementation. In contrast, our analysis highlights which aspects of person-centered documentation already exist in a real-life setting. It may therefore highlight opportunities and encourage healthcare professionals to build on existing resources and strengths.

Conclusion

Person-centered documentation is a significant challenge, but applying the concept of the discourse community can help to make it more understandable. Overall, patient perspectives were mostly absent from our sample of medical records. Established norms act as a barrier that prevents the admission of new members (patients) to the established discourse community. Our findings suggest that it is not currently feasible to provide recommendations about improving person-centered documentation, because there is little or no information on patients’ needs and wishes for co-creating shared health records. Future studies should therefore explore the patient perspective on person-centered documentation.

Declaration of interest statement

The authors have no potential competing interest.

Acknowledgments

We thank Melissa Leffler, MBA, from Edanz Group (https://en-author-services.edanzgroup.com/ac) for editing a draft of this manuscript.

Additional information

Funding

This work was supported by the Gothenburg University Centre of Person-centred Care, Gothenburg University, Gothenburg, Sweden.