A qualitative investigation into the results of a discrete choice experiment and the impact of COVID-19 on patient preferences for virtual consultations

Abstract

Objectives

To conduct a qualitative investigation on a subset of participants from a previously completed Discrete Choice Experiment (DCE) to understand why factors identified from the DCE are important, how they influenced preference for virtual consultations (VC) and how COVID-19 has influenced preference for VC.

Methods

A quota sample was recruited from participants who participated in our DCE. We specifically targeted participants who were strongly in favour of face-to-face consultations (F2F - defined as choosing all or mostly F2F in the DCE) or strongly in favour of virtual consultations (VC - defined as choosing all or mostly VC consultations in the DCE) to elicit a range of views. Interviews were conducted via telephone or videoconference, audio recorded, transcribed verbatim and uploaded into NVIVO software. A directed content analysis of transcripts was undertaken in accordance with a coding framework based on the results of the DCE and the impact of COVID-19 on preference.

Results

Eight F2F and 5 VC participants were included. Shorter appointments were less ‘worth’ travelling in for than a longer appointment and rush hour travel had an effect on whether travelling was acceptable, particularly when patients experienced pain as a result of extended journeys. Socioeconomic factors such as cost of travel, paid time off work, access to equipment and support in its use was important. Physical examinations were preferable in the clinic whereas talking therapies were acceptable over VC. Several participants commented on how VC interferes with the patient-clinician relationship. VC during COVID-19 has provided patients with the opportunity to access their care virtually without the need for travel. For some, this was extremely positive.

Conclusions

This study investigated the results of a previously completed DCE and the impact of COVID-19 on patient preferences for VC. Theoretically informative insights were gained to explain the results of the DCE. The use of VC during the COVID-19 pandemic provided opportunities to access care without the need for face-to-face social interactions. Many felt that VC would become more commonplace after the pandemic, whereas others were keen to return to F2F consultations as much as possible. This qualitative study provides additional context to the results of a previously completed DCE.

Key messages

What’s already known about this topic?

• The use of virtual consultations increased due to the COVID-19 pandemic.

• Several factors have previously been shown to influence patient preferences for virtual consultations.

What does the study add?

• This study provides theoretically informative insights to explain the results of a Discrete Choice Experiment.

• This study highlights the impact of COVID-19 on patient preferences for virtual consultations.

Introduction

The UK’s National Health Service (NHS) Long Term Plan [1] sets out a policy agenda of mainstream digitally enabled care. Virtual Consultations (VC - either a real-time phone or a video consultation) have been suggested to reduce up to a third of outpatient appointments and save ‘over £1billion a year [1]. Benefits of VC also include saving patients time, convenience and freeing up healthcare professional time. The COVID-19 pandemic has accelerated the introduction of VC into clinical practice [2] with many organizations working hard to introduce VC [3]. Technology has taken a ‘central role’ [4] in healthcare following a ‘big bang’ change in technology driven work practices [5]. COVID-19 has brought about changes in the healthcare landscape in line with policy agenda in the NHS [1].

Virtual physiotherapy has seen an increase of interest following the pandemic. Prior to COVID-19, virtual physiotherapy was to have a number of advantages, including increased flexibility, accessibility and reduced costs [6]. Digitally enhanced outpatient care has been labelled as a positive legacy of COVID-19, with the use of VC opening the door to remote working, remote assessment, remote monitoring and rehabilitation [7]. Outpatient physiotherapy services are now seen to have the opportunity to use blended digital approaches with traditional face-to-face (F2F) appointments, to suit the needs of patients, in an individualized manner [8]. Although VC was embraced during the pandemic, less than half of clinicians in a cross-sectional survey believed telehealth was as effective as F2F care [9]. The development of effective, patient centered, accessible, equitable and flexible patient care pathways has been cited as an important ambition [10]. An understanding of patient preferences is essential to the design of such innovative pathways in physiotherapy.

Preferences can be defined as a ‘total subjective comparative evaluation’ [11]. Preferences are the result of a cognitive task whereby individuals consider the alternatives and their consequences to determine the option or action which yields the greatest utility (or benefit) to them. Rational preference theory assumes that the individual will subsequently choose the option which benefits them the most [11].

The CONNECT Project [12] is series of mixed methods studies investigating patient preferences for VC and is split across four phases. In Phase 1, a systematic review was conducted that investigated how the work of being a patient influences preferences for VC [13]. Phase 2 was a qualitative study that investigated the various factors that influence preferences for VC. Phase 3 extended this work through a discrete choice experiment (DCE) [14]; a deductive investigation to test the strength of individuals characteristics and demographic factors and their relationship with preference for VC. Our previous DCE was terminated prematurely due to COVID-19 and we are therefore treating the results as indicative rather than absolute as the required number of patients were not recruited to enable definitive conclusions to be drawn. The results of the DCE suggest that people who prefer VC are: more likely to have access to the equipment required to undertake a VC and to have difficulty with activities of daily living; less likely to have resources to accommodate time and travel and to be educated to degree level. Soon after the termination of the DCE, there was an organizational restructuring to introduce virtual consultations due to COVID-19 [2] and a qualitative investigation is needed to investigate the impact of COVID-19 on preferences and provide additional context to the results of the DCE [14].

The primary objective of the present study was to conduct a qualitative investigation on a subset of DCE respondents to understand why factors identified from the DCE are important, and how they influence preference for VC. A secondary objective was to understand how COVID-19 influences preference for virtual orthopaedic rehabilitation consultations.

Methods

This research is a qualitative investigation to help us to further understand the results of phase III of the CONNECT project [14]. The CONNECT project protocol has previously been published [12].

Ethics

Ethical approval for DCE delivery was obtained for Phase III (approval received on 18 October 2019 from the London-Hampstead Research Ethics Committee - IRAS ID: 248064, REC Reference 15/LO/1586). A subsequent amendment for inclusion of qualitative interviews was granted on the 26th June 2020. All participants provided informed written consent via email prior to inclusion.

Setting

The research was conducted within a single specialist orthopaedic hospital in North London, UK. All participants were recruited from the Occupational Therapy and Physiotherapy Department.

Participants

A quota sample was recruited from participants who completed our Discrete Choice Experiment (DCE) [14]. The inclusion criteria is demonstrated in Table 1.

Table 1 Inclusion and exclusion criteria

Recruitment

Participants meeting the inclusion criteria were sent an email by the lead investigator (AWG) informing them of the research. Those who replied indicating they were interested in taking part were sent the participant information sheet. Written consent to participate in the research was gained via email. A mutually convenient time was then arranged for interview.

Data collection

Interviews were conducted via Zoom software or telephone. A topic guide, focusing on the results of the DCE, was used to facilitate discussions (see supplementary material 1). Interviews were audio recorded and transcribed verbatim.

Data analysis

Transcripts were uploaded into QSR NVIVO (version 12). A directed content analysis [16] was undertaken in accordance with the coding framework designed from the results of Phase III [14] of the CONNECT project. This took the following form:

• i. Data identified within the transcripts and allocated to the most appropriate factor group from the coding framework (pathway factors, clinical factors, socioeconomic factors, equipment factors, objective factors, interaction factors, COVID-19 impact on preference).

• ii. Data were characterised based on the question: how does this factor influence preference for virtual consultations?

• iii. The characterisation from (ii) was saved as a node within NVIVO.

Initial coding was undertaken by one author (AWG) with support from CRM. Another author (HB) subsequently reviewed all nodes within the NVIVO file to check that:

• a. Each node was an accurate representation of the interview transcript

• b. Each node fit within the coding framework.

Data were then presented with excerpts from transcripts to illustrate salient features.

Coding frame

The coding frame is shown in Table 2. We were interested in data relating to:

• i. How the context of the consultation (the circumstances of the consultation and the patient’s symptoms and activity levels) influences preference.

• ii. How patient access to resources (based on their socioeconomic position and access to technological resources) influenced preference.

• iii. How the requirements of the consultation (the objectives and whether the interactions required to fulfil the objectives) influence preference.

• iv. The impact of COVID-19 on preference for F2F or VC.

Table 2 Coding Frame

Results

Respondents

Thirty-eight participants met the inclusion criteria from the F2F group. Of these, 26 did not respond, 4 declined interview and 8 were interviewed. Seventeen participants met the inclusion from the VC group. Of these 11 did not respond, 6 consented to interview with one participant subsequently unavailable for interview. Five were subsequently interviewed. Participant characteristics are demonstrated in Table 3. Interviews lasted for an average of 50 min (range 34 to 79 min). Empirical data are demonstrated in Table 4 (Context for the consultation), Table 5 (patient’s access to resources), Table 6 (what’s required from the consultation) and Table 7 (how COVID-19 influences preference).

Table 3 demographics of participants

Table 4 Context for the consultation

Table 5 Patient’s access to resources

Table 6 What’s required from the consultation

Table 7 How COVID-19 influences preference

Context for the consultation

Pathway factors

Patients preferred virtual appointments early in the morning to avoid having to get up earlier and avoid rush hour traffic; public transport was busier during these times which was challenging for some patients and also led to patients preferring VC. Other patients however, preferred to get the appointment out of the way and were happy to travel. F2F appointments were easier later in the day as traffic volume would be reduced, there were fewer obstacles and there was a better chance of locating a parking space. Patients were less likely to prefer a F2F appointment for shorter durations, with some participants questioning whether it was ‘worth’ travelling in for only a 15-min appointment; longer appointments made travelling in more worthwhile. Some patients felt that they would rather a F2F appointment with a longer wait between sessions as ‘anything could happen’ during that space of time.

Clinical factors

Particularly for patients suffering from pain, avoidance of pain was a driver to prefer a VC. Patients who struggled with daily activities, especially getting ready in the morning, found travelling to an earlier appointment problematic. Extended travel led to an increase in pain which could last for several days and this led to some preferring VC.

Patients preferred to see a specialist F2F, particularly when symptoms were bad so that someone could physically assess them. There was a sense that VC was not suitable to address complex problems. The fear of being isolated is a motivating factor to attend consultations F2F. One participant expressed a general desire for a VC; their dislike of seeing themselves on a screen would lead them to opt for a phone rather than a video call.

F2F Face to face consultation, VC Virtual consultation, F Female, M Male

Patient’s access to resources

Socioeconomic factors

The cost of travelling to the hospital is one reason for patients wanting to have a virtual consultation, particularly if repeated appointments are required. Travelling to an appointment was more costly for patients who did not have access to a car, particularly if they needed to travel on public transport during peak travel times, which tends to have a higher cost. Taxis were particularly costly for some patients and the requirement for overnight accommodation for a F2F appointment further influenced preferences in favour of a VC. A patient’s employment was a significant factor: some could afford to take time off work to attend appointments, while others would have to take unpaid leave. These financial factors influence preferences. Patients who had a degree were assumed to be paid higher than those who did not have a degree. Participants commented on how graduate jobs may have more chance of paid leave to attend appointments. More affluent patients were able to afford to take time out of work and attend a F2F consultation.

Equipment factors

Patients who lived with or near people who could support them with accessing or using equipment were in a stronger position to be able to use VC. Those patients who have been using technology for other areas of life and were familiar with it were more likely to choose VC than those who were not. Several patients reported an increased use of technology to communicate with work or family since the onset of the COVID-19 pandemic and would now consider using VC for their rehabilitation; particularly during the pandemic. Patients who did not have access to the equipment to conduct a VC were more likely to prefer a F2F consultation. In addition, poor internet connectivity was off-putting to patients.

What’s required from the consultation

Objective factors

Respondents expressed they were happy to have a virtual consultation if a physical examination was not required. Participants were happy, in general, to have a VC for a discussion. It was recognized that a fluctuating condition might require different input at different times. Basic rehabilitation was acceptable to some, others preferred any form of rehabilitation to be carried out in person. First appointments were generally seen as better if they were conducted F2F, particularly if physical rehabilitation was required to ensure exercises were being completed correctly. Follow up appointments were deemed to be more acceptable via VC, particularly if the clinician was known to the patient. If an issue required a thorough assessment F2F was identified as the best option.

Interaction factors

One participant in the F2F group argued for first appointments to be conducted virtually to allow for a (subjective) assessment to be conducted to plan care. This was at odds with most of our DCE respondents who preferred F2F for their first appointment; this demonstrates the individual nature of preferences. Virtual care was best with a therapist who was known and trusted by the patient, with a good rapport facilitating preferences in favour of VC. Participants who reported travel to be a challenge described previous experiences where their interactions with healthcare professionals were inhibited by symptoms, such as the inability to focus on the content due to pain. It was thought that, for those patients who do not speak English, accessing a family member to support translation would be easier from home. Participants who had established relationships with their clinicians had confidence in VC. Interactions were better F2F rather than VC as it was easier to see body language. One participant referred to interactions as ‘cold’ virtually [3BF05] and several commented on how VC created the illusion of clinicians not listening as intently and potentially becoming distracted. Physically attending gave the potential for more empathy which was important. Intimate examinations over VC might make patients feel uncomfortable which may influence interactions.

How COVID-19 influences preferences

VC during COVID-19 has provided patients with the opportunity to access their care virtually without the need for travel. For some, this was extremely positive. The pandemic highlighted the potential use of VC technologies and participants in this study thought that their use has increased across society. The potential benefits of VC in healthcare have become apparent to participants whereas these benefits were not previously visible. The healthcare and pandemic situation is different for the participants in this present study compared to when they completed the DCE (pre-pandemic). Due to this, participants stated they would answer the DCE differently if it were to be undertaken during the pandemic.

Participants were fearful of catching COVID-19 and could see that VC offers an opportunity to access care without being put under any undue risk of transmission. Travel, particularly on public transport, was seen as a high-risk activity for patients and some participants stated they would avoid this wherever they could. COVID-19 influenced patients’ preferences; many rationalised the trade-offs between travel and virtual care and although they would normally prefer F2F they would, under the current circumstances, opt for a VC. Despite this, a small number of participants expressed they would still travel in for their rehabilitation if this were available.

A hospital environment was viewed as a sterile, clean, place where there would be low risk of COVID-19 transmission. Participants cited infection control policies and procedures and would be happy to travel if they had access to their own transport. One participant suggested waiting in the car park until the clinician was ready to avoid spending unnecessary time within the hospital. Clinicians wearing Personal Protective Equipment inspired trust and one participant commented how they felt they were more likely to contract COVID-19 in a supermarket.

Although participants who were strongly in favour of F2F prior to the pandemic would consider undertaking a VC during COVID-19, they expressed a continued desire to have F2F consultations after the pandemic. Despite this, a greater appreciation of the potential benefits of VC was felt by all participants. It was felt by some participants that once the pandemic was over F2F care will become the norm once again.

Discussion

Despite the DCE being terminated prematurely due to the COVID-19 pandemic, the results from the DCE suggested a tendency for certain patient groups to have preferences for VC [14]. A sub-sample group of participants with strong preferences for and against VC were identified from the DCE to participate in this present study. This study investigated the results of our previous DCE and provides additional useful insights. Thirteen participants (8 strongly in favour of F2F, 5 strongly in favour of VC) were interviewed to investigate the results of our DCE. In addition, several reasons why COVID-19 may have changed patients’ preferences towards VC during the pandemic were identified.

Our DCE [14] indicated that patients preferred VC when the therapist was known to the patient, there was a longer time until the next appointment, a shorter appointment early or late in the day; for patients without a degree, who had access to the equipment they need, had difficulty with day to day activities, were undergoing rehabilitation for multiple problem areas and hade to pay more than £5 for their return journey. Conversely, the opposite pre-conditions (when the therapist was not known to the patient, a shorter time until the next appointment, a longer appointment, in the middle of the day; for patients with a degree or above, did not have access to the equipment to undertake a virtual consultation, did not have difficulties with day-to-day activities, were undergoing rehabilitation for a single problem area and had less than £5 to pay for their return journey) led to patients preferring a F2F.

Patients preferred not to travel early in the morning for therapy if they had difficulty getting ready, had to wake up unacceptably early or did not like driving in rush hour. Ackerman and colleagues [17] identified that patients had preferences for certain times of day and this being a reason for not attending a self-management course. We have demonstrated how time of day can motivate preferences, with appointments in the middle of the day being easier for some due to reduced traffic and easier parking. Public transport can be more costly at peak times (i.e. early morning). Patients with musculoskeletal conditions may experience morning pain and stiffness [18]; these morning symptoms contributed towards patients preferring VC for an early appointment. Some patients may appreciate being able to spend time and energy gained from not travelling on other activities [19]. Elimination of transport time when using VC has been shown to be a significant benefit [15]. Our recent systematic review and qualitative synthesis [13] highlighted how changes in the work of being a patient influences preferences; if factors relating to travel and logistics make the work more burdensome for patients they are more likely to prefer an option that is less burdensome. PhysioDirect [20], a randomised trial investigating telephone advice and assessment services for physiotherapy, was more successful when calls were made at a convenient time for patients [21]. Time of appointment may not be a true reflection of preference for VC, rather the option of VC at that time making an appointment more convenient for patients at that time. Some patients in this study, however, liked an earlier appointment so they could travel in and get to work or other commitments earlier in the day. Some patients have reported the benefits of undergoing a Skype consultation from work [22]. Trends identified by our DCE do not apply to all, preferences are clearly individualised.

Being able to take paid time off work was important to allow F2F attendance with reduced financial burden. People in education to school leaving age are over represented in ‘zero hours’ contracts [23] and therefore may be unable to take paid leave for medical appointments. This may be challenging for some who have undergone surgery before their rehabilitation and been forced to take time off work previously [24]. A participant in our previous research [25] described how appointments had become a full time job; repeated attendance can get in the way of employment and travel can be financially demanding.

Equipment can be costly; ‘Attend Anywhere’, the platform of choice across the NHS in England and Scotland, requires Windows 7+, MacOS 10.11+ (released 2009 and 2015 on Windows and Mac respectively) on a desktop and Android 5.1+, iOS 12.4+ (released 2015 and 2019 on Android and Apple phones or tablets respectively) with up to date Chrome or Safari software [26]. A patient’s financial position may remove the opportunity for VC through the initial purchase and the ongoing costs of some software that drive up data usage costs. When outdated hardware was incompatible with the platform, this led to reduced patient satisfaction [2].

Many of the patients in this study preferred to have a F2F prior to a VC, although one patient reported they would be happy for an initial assessment. Other studies [15, 22] reveal how patients favoured initial F2F appointments prior to VC. For our patients, this was to conduct a thorough assessment and to learn the correct exercises. If a patient was experiencing a worsening of symptoms, they are more likely to want a F2F. VC offers flexibility [27] but patients might also want a F2F to identify the cause of a new problem should it arise [22]. Some patients felt that VC would not be accurate whereas Cottrell and colleagues [28] found high levels of agreement between in-person assessment of patients and telehealth appointments. Teleconference goniometry has been shown to be as accurate as in person goniometry of the elbow [29].

Therapists are forced to rely on their talking and listening skills (as opposed to hands on) which may be problematic for patients with communication difficulties; ordinary conversation has been demonstrated to be a key factor of a therapeutic relationship [30]. One participant in this study reported not liking seeing themselves on a screen, participants with social anxiety disorder have been shown to have self-focused attention during conversations using Skype [31]. Patients have been shown to be skeptical about telephone appointments prior to use [21, 32]. However, trying out a VC platform has been shown to increase positivity about ease of use and usefulness compared to those who did not use it [15] and may alter the perception of the patient-therapist relationship [15].

Nationally in orthopaedics, routine care was abandoned due to COVID-19 to reduce patient ‘flow’ to prevent the NHS being overwhelmed [33]. The NHS now faces an estimation of 400,000 procedures not being performed every month [34]. Virtual orthopaedic consultations have subsequently been hastened and rapidly implemented [2], with new guidance for virtual care being disseminated widely to support use in orthopaedics [35]. Patients in this study indicated that their stated preferences in our Discrete Choice Experiment (conducted between December 2019 and March 2020) would have been more favourable towards VC if they were able to foresee the impact of the pandemic. Patients did not feel F2F was viable during the height of the pandemic and were not happy to take public transport. Patients were using platforms like Zoom and WhatsApp to communicate with friends and family and reported they felt more confident with using VC to access care; prior to the pandemic some patients had not used these technologies. Using VC highlighted the benefits of not travelling and saving money. Some patients, however, would still be happy to travel for a F2F appointment as they believed the risks of transmission would be low with proper precautions. It was suggested by some that everything would return to normal after the pandemic subsided and F2F would resume once more. Interestingly, of those patients who were unable to have a F2F due to COVID-19, less than half of VC patients would prefer a VC next time [2]. The pandemic has affected preferences in the short term, what is not clear is how preferences will be affected in the long term.

Strengths and limitations of this study

Our previous DCE results indicated factors that influenced preference for VC or F2F. The interview schedule and coding frame from this present study reflected this, and it is possible that different or additional questions may have yielded different results. Of the participants eligible for inclusion, 68% did not respond to the initial or follow-up email; an increased number of participants may have changed our conclusions. The limited pool of participants who had strong preferences for VC reduced our potential sample and as a result the recruitment reflects a larger number of participants in favour of F2F from our DCE. An alternative sampling strategy may have led to a higher level of recruitment than was observed in this study. Higher recruitment numbers may have influenced our conclusions. Despite these limitations, we have been able to sample groups of patients who were able to offer a diverse range of perspectives. We have used theoretically informed qualitative methods to interpret a DCE through interviewing these participants to understand what they think these results mean. These results will be of particular interest to the physiotherapy and rehabilitation community who are using virtually supported consultations in their patient pathways.

Conclusion

This paper presents a study that investigated the results of a discrete choice experiment and has explored the impact of COVID-19 on patient preferences for VC. Patients suggested a range of potential reasons as to how the context of the consultation, patient’s access to resources and the requirements of the consultation might impact their preference. In addition, patients shared experience and viewpoints on how the COVID-19 pandemic has influenced preferences for VC. VC during COVID-19 has provided patients with the opportunity to access their care virtually without the need for travel. For some, this was extremely positive as it provided opportunities to access care without the need for F2F social interactions and potentially risk contracting the virus. Many felt that VC would become more commonplace after the pandemic whereas others were keen to return to F2F consultations as much as possible. This research sheds light on some of the underlying rationale behind patient preferences for VC in certain situations.