We won’t stop Googling our symptoms any time soon, so why aren’t online health information resources more effective? An interdisciplinary study between Computer Science and Health Sciences is trying to answer the question, explains Julia Mueller.

The internet is increasingly becoming an important health information source and many people routinely turn to the internet to diagnose symptoms. This could be important for illnesses like lung cancer, where symptoms are often ignored for a long time.

We were interested in what lung cancer patients and their family members report about the time before the diagnosis and what information they found helpful to appraise symptoms. We conducted a survey and interview study among recently diagnosed patients and asked them about their experiences. The participants reported using the internet to decide whether seeing a GP was necessary. They also used it to challenge their doctors’ advice when they felt dismissed, or not taken seriously enough.

In addition, they went online to look up medical jargon to help them to understand why diagnostic tests were done and what the results meant.  For example, the daughter of a patient told us of her frustration when her mother was told repeatedly by her GP that her symptoms were merely due to an allergic reaction. After several visits to the GP that yielded no improvement in the symptoms, the daughter researched the symptoms and then requested that further tests be done. She believed this was the trigger for further investigation, and found the Internet helpful in this process. She also referred to the Internet whenever the doctors used jargon she didn’t understand to describe the test results, like ‘pleural effusion’. This helped her prepare for a potential cancer diagnosis and helped her cope.

Based on our results, we developed some recommendations for health websites and how these can encourage people to see their GPs when they have symptoms related to lung cancer. To test these recommendations, we created our own lung cancer advice website with a symptom checker tool called ‘When2Go’.

On entry to the When2Go website, people are asked to input some information about themselves and their symptoms. They then receive some personalised information on symptoms, risk factors and NHS guidelines on when to see a GP. Currently available cancer websites generally provide a list of symptoms which patients look at and decide for themselves whether to visit their GP. The new web tool, When2Go, aims to provide more personalised advice by asking questions about symptoms such as unintentional weight loss, or persistent coughing. As people are asked, and answer, more questions the software builds up a more accurate picture that culminates in a recommendation of whether or not to visit a GP. It also generates a report which the user can bring with them to their GP appointment.

We are comparing this tool against other versions – such as a generic, impersonal version – to test which components of the tool are effective. Overall, the tool works by supplying advice on when to seek help and which symptoms should be examined by a GP, rather than promoting self-diagnosis. It can’t tell you whether you’ve got lung cancer, but it can tell you when it’s best to see your GP.

With this study, we hope to identify concrete ways in which health websites can improve their content to encourage presentation to health services when it is necessary. Many people (including researchers and health professionals) are sceptical about looking up symptoms online and argue that this can lead to unnecessary panic and worry.

Research looking at log data from search engines has shown that people often progress from researching mild, common symptoms to rare, serious diseases – even though benign causes would be much more likely. For example, people often progress from researching a common symptom like ‘headache’ to a serious but rare cause such as ‘brain tumour’. This is despite benign explanations like caffeine withdrawal or lack of sleep being much more likely causes. This is termed ‘escalation’ and has been related to hypochondria, sometimes even dubbed ‘cyberchondria’. This raises concerns about the impacts of online health information on the demand for health services and whether it causes people to over-run GP practices with common symptoms, unnecessarily wasting NHS resources.

There is another side to the coin, however. In June 2015, a news article reported the case of a young girl whose cancer recurrence went undiagnosed for a prolonged time “because doctors, specialists and GPs at Nottingham University Hospitals Trust failed to listen or believe her complaints”. The girl had researched her symptoms online and had come to the conclusion that her cancer had recurred, but “her increasingly urgent pleas for help were met with an admonishment to ‘stop Googling’ her symptoms”.

This case shows how online health information could help people appraise their own symptoms and facilitate earlier diagnosis. But this requires health professionals to be prepared to accept empowered patients who want to take active, informed roles in the diagnostic process.

So while I agree with scepticism towards online symptom research, research suggests the internet is becoming increasingly important as a health information source and can play a positive role. People won’t stop Googling their symptoms, so we need to make sure they can access helpful information when they do.

Health websites need to take psychological theories about behaviour into account to tackle psychological reactions to symptoms, such as fear, denial and embarrassment. By building-up a detailed picture of participants in this trial we hope to make the advice better. Anything that stops people with symptoms putting off seeing their doctor could help to save their lives.