This thought piece discusses the requirements for development in healthcare technology needed to support the improved management of multimorbidity within modern healthcare systems. It is written from the perspective of a General Practitioner with experience in the development of healthcare software and solutions.

Defining Multimorbidity

Multimorbidity is defined by The National Institute for Health and Care Excellence (NICE) as the presence of two or more long-term health conditions, this can include:

  • Defined physical or mental health conditions of long duration, such as diabetes, schizophrenia, or an infectious disease.
  • Ongoing conditions, such as learning disability.
  • Symptom complexes, such as frailty or chronic pain.
  • Sensory impairment, such as sight or hearing loss.
  • Alcohol or substance misuse.

Globally, the prevalence of multimorbidity is approximately 30% — this increases significantly with age and socioeconomic deprivation and is almost universal in older adults. It is important as it is associated with:

  • Reduced quality of life.
  • Reduced life expectancy.
  • High treatment burden.
  • Mental health difficulties, such as anxiety and depression.
  • Fragmentation and problems of coordination of care.
  • Functional difficulties.
  • Polypharmacy and higher rates of adverse drug events.
  • Increased use of health services (including unplanned, or emergency care).
  • Poorer quality of life for families/carers.

The potential renaissance of General Practice

When NICE released its guidance on multimorbidity in 2016, I felt that it had defined a major workstream for General Practice. The guidance built upon the strengths of General Practitioners suggesting that people with several long-term conditions needed to have discussions with their GPs to develop appropriate “bespoke” plans for managing their conditions. These plans may vary, even for individuals with apparently similar illness profiles as they would reflect their priorities and preferences. For me, this represented true personalised medicine with plans crafted and agreed by people with their trusty GP. This had elements of my Royal College of General Practitioner membership exam where I would consider the patient in the context of their physiological, psychological, and social circumstances and engage with them to explore their concerns, ideas, and expectations.

NICE Guidance

Guidance entreated clinicians to consider the treatment burden of multiple medications, to stop medications where possible and to have risk-based discussions with our patients. Whilst NICE were pushing against an open door with respect to my heart and mind, implementing the guidance gave me some challenges. At the time my principal role was working in software and services; I still performed GP sessions but not in a pattern that gave patients the continuity of care that this approach may require. There were other challenges facing most GPs though. Firstly, as patients went to specialist clinics where the competence and even sanity of their GP might be questioned for failing to follow the precise guidance for the condition of the day. Secondly, having risk-based discussions can be difficult without the ability to present evidence to patients in a manner that they can digest across several conditions.

System Wide Care Plans

As a clinician working in technology my first thought was that we needed to be able to capture the agreed management plan following discussions within some system wide, integrated plan visible to clinicians working in primary, community, pre-hospital, emergency, inpatient, and outpatient care. This transcends mere interoperability of medical records which remains a modest aim for many systems. We are seeing major improvements in information sharing within UK with initiatives such as the Yorkshire Care Record and similar developments in London and Greater Manchester. In the latter development heart failure care plans are shared across the system but how do we “scale up” plans to cover, for example, a patient with heart failure, COPD, and type 2 diabetes?

Supporting Risk Based Discussions

My next requirement is more complex, how do we explain to people why they might consider starting, continuing, stopping, or avoiding medications? This is a more complex version of the “statin conversation” that will be familiar to most if not all GPs. Where individual care pathways demand specific treatments, we must be careful to avoid layering these treatments slavishly to the detriment of the patient. Risk-based conversations in primary care generally link to risk calculators and numbers needed to treat. They become complicated discussions for someone with several long-term conditions.

My personal approach where medication is not directly relieving symptoms is to talk about insurance policies, most people understand this and quickly grasp that, no matter how comprehensive a policy is with promises of cover, if you cannot afford the premiums, it is not for you. On a day-to-day basis the “premiums” in this care are side effects and interactions. An additional “premium” is the sheer burden of medications that people must manage.

But consider the complexity of the risk discussion where, for example, we are consulting with our patient with heart failure, COPD, and type 2 diabetes. My belief is that this is where creative use of artificial intelligence and machine learning may provide supportive solutions. In this case we are not seeking a single answer but an ability to provide evidence in a digestible way so that we can sort out the “no brainer” interventions from those that might be considered.

Agile Patient Management

My final reflection on this piece is an insight not about technology per se but about the methodology used to develop it. Agile project management has, I believe, something to teach us clinicians. Agile projects have a backlog, effectively everything that can be done this is kept under review and is “groomed” to refine it, it is impossible to progress everything on the backlog so individual elements are prioritised and selected for a development “sprint” at the end of the sprint period there is a progress review, and the next sprint is planned. This next sprint may contain unfinished items from the previous sprint to which new prioritised items are added. For me, this is a classic structure for managing patients with multimorbidity. We don’t have to cover everything in a single consultation or time period just the most important issues or those that can be opportunely addressed. We then schedule a review to assess progress before planning the next “sprint”. This clearly requires a degree of continuity of care but the role of technology in supporting continuity is for another blog.