General Practitioners face a daily attack on their cognitive and decision-making skills. Mental fatigue is shown to affect physical performance and increase perceived exertion (Van Cutsem 2017).

The trend in healthcare is for increasing amounts of data: evidence-based medicine, electronic health records (EHR), clinical guidelines, personal monitoring, test results and reports. The consultation in Primary Care is often truncated into 15 minutes and can be prone to interruption. Clinical contact generates an even larger amount of non-clinical work. All this makes a difficult job, exhausting.

Medicine, especially General Practice, is about managing complexity and risk. As evidence-based medicine expands and data increases, so does the complexity of Medicine. Complexity and risk are directly associated; increasing complexity, increases risk.

Employing mental shortcuts, or heuristics, allow Clinicians to manage the complexity of Medicine and the clinical case before them. Pattern recognition, previous experience and knowledge of guidelines allow steps to be skipped and mental energy saved. Taking all factors into account, making balanced, reasoned arguments and producing a diagnosis and management plan, for every patient would require vast amounts of time and energy.

However, despite heuristics, General Practice is not coping, the situation is getting worse. Real exploration needs to be undertaken as to why we are not reducing the mental burden on Clinicians. Much like the aviation industry, we need to start looking at the health and safety of our pilots and crew.

Experientially, patients appear to be getting more complex in Primary Care, often with multiple co-morbid medical problems. The drive to bring care close to a patient’s home also means the drive of medical conditions, that previously sat firmly in the domain of secondary care, being managed by the generalist in a Primary Care, rather than the specialist in the Hospital.

If Primary Care, and by extension General Practice, is to survive, there is a pressing need to remove the noise associated with working in Medicine. We need to reduce the ‘paperwork’ and allow Clinicians to focus once again on practising Medicine.

Informal data collection after discussions with colleagues on the problems identified in their daily work, has provided me with ideas for five areas to explore in reduce GP workload:

The digital urgent admission

The pain of admitting to a patient to the on-call team is a pain known only too well in Primary Care. You can be running to time one minute and after 15, 20, 30 minutes on the phone waiting for the appropriate clinician to answer, your surgery becomes overly pressured, the thought of being late home, the reflection of ‘why is it so difficult?’ keeps returning to you brain in following consultations. All this contributes to fatigue.

Primary Care for a long time has been at the forefront of technological advances in the NHS. There is a feeling now that we are waiting for the infrastructure, for other areas to catch up. What if we don’t have to? What if we can employ technology, that makes GP easier, but can still integrate with some of the archaic elements of the NHS.

I give you the new urgent admission referral…

Ok, so this was a hair appointment. The question is, how far are we from an assistant that can call the admitting team, convey relevant information from the notes you have just written and allow you as a clinician to intervene if further information is required?

All it needs now is a ‘surliness factor’ that can be increased for those particularly difficult Friday afternoons.

Automate the mundane

There are multiple tasks and workflows in General Practice that are repeated ad nauseam. Writing in medical notes, transferring information into a dictated letter, reviewing test results and needing to contact or review a patient, completing form for XYZ, completing medical report forms for insurance, medicals etc.

For every hour of patient contact a further two hours are required interacting with the EHR (Sinsky 2016). Want a more productive GP, reduce the mundane and the repetitive. If I want to refer a patient, let me click a button that pulls the data and sends it to the hospital, don’t make me dictate a letter and duplicate my work. Need a medical report? Let the system generate a report, that I can read, check and send. It is time for the EHRs that we work with to start paying back, we have put so much into them, it is time for them to deliver on their return.

Remote monitoring

Remote monitoring brings unique insights into the health and lifestyle of our patients. We can monitor disease and ‘wellness’ without having to contact the patient. The flip side is we have access to unprecedented amounts of information about patients outside of the GP surgery. Putting to one side the legal implications and who is responsible for abnormal results, this represents a vast amount of information, for which the majority will be absolutely fine.

Remote monitoring is here, I can track my steps, pulse, exercise frequency, sleep without even considering it. The real question for Primary Care is how we deal with this data? How do we transform data into information, into knowledge? Personally, I get a slight sadness whenever a 7 day home blood pressure monitoring form finds it way to me. As I sit there working out the average systolic and diastolic blood pressure, I think ‘am I the only one doing this? Am I the only one who it upsets?’

My point is a simple one, using a GP to analyse data by hand is pointless and a waste of money and resources. The real advance is in the development of algorithms to process this data.

Use whatever technique you want, Machine Learning if it makes you happy, but the point is, as a GP, the only interaction I want to have with remote monitoring is when there is a patient in front of me because the device has predicted a decline in health or that the person is at risk of a disease. I don’t really need to see the numbers, it would be nice to have the ability.

Time to analyse our EHR

Electronic Health Records (EHR) have a vast amount of information. Machine learning has been employed to predict what differential diagnosis or medication will be required for a patient’s next visit (Choi 2016). The EHR has a wealth of information; a simple algorithm to create would be predicting time to next visit for a patient based on diagnosis, differential diagnosis or lack thereof.

Continuity of care is not only better for patients (Isaac Barker, 2017), but results in less ‘cold-start’ consultations, i.e. consultations at the beginning of a problem with no prior knowledge of the patient, or worse, becoming involved midway through a disease. More frequent contact with the patient’s regular practitioner can result in 8.96% to 12.49% fewer admissions, depending on whether there was a medium or high level of continuity, respectively (Isaac Barker, 2017).

‘Cold-start’ consultations require greater cognitive effort on the part of the clinician and, from personal experience, the patient often tries to summarise their symptoms and what has been happening to speed up the consultation. Unfortunately, summarising often leaves out vital information resulting in repetition of tests, incomplete differential diagnoses and poorer management.

What if we could predict when a patient may represent, based on previous evidence? Could we pre-emptively book an appointment for a review, thus seeing the same clinician?

Integrate computer systems

New technology is coming into healthcare at an amazing rate, but it is not without its problems. New systems, designed to improve patient care or user experience, if not integrated properly can result in increased work for the user.

As an example, in the GP system EMIS, tasks can be sent regarding a patient. Docman, the letter and document management application can also create tasks. Emails can also generate tasks. There are three systems all generating work, there is an inherent risk of missing a task because it is in one environment and not another. There is also the problem that checking 3 systems recurrently through the day generates additional work.

Imagine a system that is fully integrated, all tasks, all patient information is in one place and not divided amongst different systems.

Conclusion

It’s time to think smart about General Practice. There is a wealth of technology already in the NHS, but lets focus on how we can use it and expand it to solve problems. New technology must be designed to solve a problem, we must not focus on solutions that need a problem.

References

Van Cutsem J, Marcora S, et al. The Effects of Mental Fatigue on Physical Performance: A Systematic Review. Sports Med. 2017 Aug;47(8):1569-1588.

Choi E, Bahadori MT et al. Doctor AI: Predicting Clinical Events via Recurrent Neural Networks. JMLR Workshop Conf Proc. 2016 Aug;56:301-318. Epub 2016 Dec 10.

Isaac Barker, A. S. Association between continuity of care in general practice and hospital admissions for ambulatory care sensitive conditions: cross sectional study of routinely collected, person level data. BMJ 2017

Sinsky C , Colligan L , Li L , et al . Allocation of physician time in ambulatory practice: a time and motion study in 4 specialties. Ann Intern Med 2016;165:753–60