The healthcare industry is notoriously bad at communicating with patients. It’s thought poor communication could cost the NHS a billion pounds each year. Better communication with patients has been associated time and again with improved outcomes – a recent report by Marie Curie quotes a multitude of studies proving this.
Non-adherence to drug programmes is one of the consequences of ineffective communication between the healthcare industry and its patients. In the US, it’s thought this could cost $100bn per year. In the UK, a third of medicine-related adult hospital admissions are thought to be the result of non-adherence to drug regimes.
Round the world, there have been many attempts at improving communications with patients. In Pakistan, high rates of patient illiteracy mean that pictures and symbols are sometimes used to give instructions on how to take medication.
Illiteracy isn’t just a problem in poorer countries. The pharmaceutical industry doesn’t really have a solution for the one in five British adults who are considered to be functionally illiterate or the one in twenty that has the reading age of a five-year-old.
It’s difficult to meet labelling requirements and also communicate effectively with patients of all reading levels. In tightly regulated markets such as the UK and US, medication tends to come with lengthy paperwork describing its proper use and giving lots of information about the product.
This excess of information isn’t necessarily digestible to the patient. Common complaints include the print being too small, or the information hard to understand by a layman. According to the Royal National Institute of Blind People (RNIB), one in five people in the UK can’t read the medicine labels because the font is too small.
Medication labelling isn’t the only area where patients often struggle with the information they are given. There have been some schemes to give patients access to their own health records in the interests of transparency.
Where patients have access to their own health data, there’s a lot of room for misunderstanding. Health records tend to be written by physicians, for physicians, and whilst transparency is important it’s hard for patients to interpret their own medical records without an intermediary.
Doctors involved with trial schemes have tended to be resistant out of concerns patients may misinterpret their own information. The main reason for doctors to resist the idea of patients having access to their records is concerns that they may then burden doctors with excessive questions.
What’s needed is a way to interpret this information to patients, without necessarily burdening frontline health workers with additional work.
An AI solution?
It’s clear that the healthcare industry needs to find a way to communicate effectively with patients in a way those outside the medical profession can understand, without compromising their regulatory obligations or burdening front-line workers.
One tool that could help the industry is natural language processing, a kind of artificial intelligence that’s good for interacting with people and answering their queries in either written or spoken form.
NLP could be used to interpret complex medical information and render it into layman’s language. Initial experiments show that patients tend to need support understanding compound medical terms even if they understand the individual component words in terms such as ‘community-acquired pneumonia’ or ‘post-surgical stage’.
Self-teaching computer intelligence programmes may be able to improve their performance and get better at how they interpret medical information based on feedback they get from users.
What’s particularly helpful is that NLP may be more approachable than human doctors for patients embarrassed by their ignorance, or facing conditions they feel awkward discussing.
Whether used in a live clinical setting or in an online environment, it may help patients discover more about their medical situation. There’s even hope that it could lead to better outcomes if it empowers patients to feel more involved in their health management.
Of course, there are many barriers to adoption of NLP in healthcare interpretation. One key barrier is the fact that patient health data often isn’t available digitally. But digital data would certainly be available for medications. There’s definite potential for NLP helping patients understand the labelling on their medication, and perhaps for helping enhance their understanding of their condition.
Whilst it’s early days yet for the use of NLP to interpret health information, there’s significant potential to use this type of technology to promote patient health.
Perhaps the most significant change that NLP offers is that human healthcare providers would no longer be the only option for patients seeking information about managing their health.
That’s a fairly radical change to the existing healthcare setup. In a few years time, NLP could make a huge difference to the way patients interact with health services. There’s room to be optimistic about the effect of this technology on health outcomes if it improves patient understanding.