This week, the British Medical Journal published the first results from one of the world’s most complex randomised controlled trials. Researchers at the Nuffield Trust led this analysis, which relied on collecting over a billion records of administrative data from more than 250 health and social care organisations.
The trial’s aim was to evaluate “telehealth” – a way of using technology to support people with long-term health conditions such as diabetes, heart failure or chronic obstructive pulmonary disease. Telehealth devices allow patients to measure things like blood sugar levels in their own homes and then transmit this information to health care professionals working remotely.
Advocates claim that telehealth helps patients manage their conditions and enables faster response from professionals in the event of deteriorations in health. This means better quality care that can potentially prevent expensive hospital admissions.
The results published this week concern the impact of telehealth on use of hospitals and mortality.
In some respects, the findings appear impressive: there were 20 per cent fewer emergency hospital admissions among telehealth patients than controls. But detailed analysis reveals that the reductions were from a low base.
While control patients experienced an average of 0.68 emergency admissions per head over twelve months, telehealth patients had 0.54. This is a difference of 0.14, or around one-seventh of an emergency admission per person.
There are reasons to be cautious. For example, emergency admissions appeared to increase among control patients shortly after the trial started. This raises questions about whether the differences seen in emergency admissions were caused by the processes used in the trial rather than by telehealth.
Investment in telehealth is often justified on the basis that it will reduce hospital cost. However, in this trial, the reductions in hospital admissions translated into only modest reductions in costs, of around £188 per person based on the amount that NHS commissioners are obliged to pay for hospital care.
These differences were not statistically significant, so could have been the result of chance. Against this, the NHS will have to balance the cost of the intervention, which is yet to be published.
Aside from costs, the other key finding for this part of the study concerned mortality. Fewer telehealth patients than controls had died by 12 months (4.6 per cent of patients compared with 8.3 per cent).
As another study has found increases in mortality rates due to telehealth, more research will be needed, as the impact of telehealth may depend on how it is implemented.
The first results are only one part of the evaluation of the Whole System Demonstrator trial. Other themes, undertaken by evaluation partners, address the impact of telehealth on quality of life, cost effectiveness and individual experiences, as well as the organisational factors relevant to implementation.
Already, the findings from the trial are proving to be complex. Commissioners will need to take the whole set of findings into account when making decisions about telehealth.