Context, scope and priorities

The observations from our case studies show that it would be very difficult to conduct this type of telehealth outpatient consultation in settings other than at the tertiary hospital for the specialist and at a suitably equipped healthcare facility for the patient. In this section I explore the implications of context for tertiary outpatient telehealth. Contexts

I observed two levels of context being implicitly referred to during the consultations of our pilot trial:

• The broader hospital context. The patients were keenly aware of this context and it shaped their demeanour and behaviour. The building and its procedures were familiar to them and were invested with a level of trust by the patients and their families. This context contained other services, such as radiology,

complex treatment was needed it contained other specialists. This context created pressures on the patient and family: getting to the next appointment within the hospital, dealing with tired children (especially younger siblings of the patient), even rushing to catch an expiring parking meter.

• The patient’s clinical context. This context resided in the treatment history of the patient, their relationship with their surgeon, their medical records, and the memories of previous inpatient and outpatient events during their treatment. These contexts influenced the evaluation responses from the parents and patients, and created potentially confusing or contradictory responses for the researchers to analyse. Two examples illustrate this point:

• One family responded that they would (hypothetically) have been happy to have had this outpatient consultation using a telehealth mode at a location more convenient to their home. Further into the exit interview they stated that it was important to them that the face-to-face phase of the consultation occurred. Their responses to the more open questions in the interview showed that they had other appointments at the hospital that day and, for that particular day, it was important for them to have come into the hospital.

• A second family gave the same pair of responses, but their explanation was that their young child had formed a strong friendship with the surgeon and would have been disappointed if he had not seen the surgeon in person.

For the researchers to adequately understand and interpret the Likert Scale and yes/no responses from the participants they need to be aware of these wider contexts.

Scope of the pilot trial and priorities for the clinicians

Our pilot trial was located entirely within the hospital, with the patients’ and surgeon’s rooms set up in two meeting rooms on the same floor of one wing of the hospital. The activities of the pilot trial concerned a very specific component of the overall treatment of the patients who had volunteered to take part in the trial. The scope of the pilot trial covered some, but not all, of the activities that were required for the consultations that were scheduled for the trial and this concept of “scope” is important in understanding and interpreting the trial.

Examples of items which were in scope:

• Voice and gesture communication, shared access to video and image data on the tablet displays, the surgeon directing the assistant to examine the patient

• The standard consultation structure: greetings, discussion of the recent past, raising and addressing issues, examining the patient, reviewing radiology data, proposing and agreeing on a plan for the immediate future, closure

• Patients presenting for review appointments after surgery Examples of items which were out of scope:

• Patients presenting with new symptoms

• Shared access to the patient’s medical records between the surgeon’s and patient’s rooms

• The surgeon writing follow-up referrals (X-ray, physiotherapy and other allied health treatments) for the patient

• A surgeon needing a viewpoint which was not supported by the existing video cameras

When something occurred which was out of scope the evaluation process needed to deal with it. A more mechanistic evaluation process, for example recording success or failure, might have flagged these out-of-scope events as failures, thus masking the underlying causes. A human-centred approach of the style used in these case studies would be aware of the boundaries of the trial and would note the out-of-scope event as a valuable piece of feedback for the next iteration of development of the system under trial.

The flow of events, and in particular those events driven by the clinicians’ priorities in looking after the interests of their patients, sometimes masked the presence of these out- of-scope events. This point is illustrated by some examples:

• One patient presented with new symptoms of the same underlying condition but in different limbs, and so there was no X-ray data of those limbs. There was also some difficulty showing the surgeon an external view of the extent of the problem. The surgeon stepped out of the role of the trial by sending her back to Radiology for additional X-rays and scheduling a second appointment later that day. From the clinicians’ and patient’s points of view this was an acceptable outcome.

From a mechanistic evaluation approach the “tick the box” response to the question “would you have been satisfied to let the patient go home after just the telehealth phase” would have been “No”. From a human-centred evaluation approach it highlighted that people matching patients to opportunities for a telehealth appointment might not be 100% correct because they may not have the most up-to-date information.

• The medical records at this hospital exist in paper form and could not be meaningfully shared between the surgeon and the assistant. The surgeons dealt with this by modifying their normal procedures after the first week and briefing the assistant on the case in the few minutes before the patient arrived.

• The surgeons chose to offer but not to write actual referrals during the telehealth phase. Instead, they wrote them during the face-to-face phase, retrieving the patient’s bar-coded labels from their medical records as necessary.

Future researchers might try to design a trial protocol to handle these events but my observations suggest that the clinical complexity of this tertiary level of healthcare creates flows of the consultations that are all subtly different. Out-of-scope events are likely to occur and the protocol may not be able to deal with them.