Features

Turning data into useful knowledge to improve patient care

Making decisions based on poor data, a concept known as ‘garbage in, garbage out’, unfortunately remains part of healthcare today, according to Alison Leary

Making decisions based on poor data, a concept known as ‘garbage in, garbage out’, unfortunately remains part of healthcare today, according to Alison Leary


Picture: iStock

Every time you shop, take public transport, use energy, access the internet through a search engine or send an email, you leave an invisible footprint in the form of a data trail.

Many organisations collect this data to see how you shop, what you buy, how you behave and if other factors, such as the weather, affect your decisions. They look for patterns in this data and use it to predict how you will shop in future and how they can nudge you into buying a little bit more, or to ensure their logistics chain can deliver goods in time. The data you leave behind is an important feature of helping others understand how you live.

The value of data

It is said that data is becoming the ‘oil’ of the 21st century as we develop a data-driven economy (Rotella 2012). However, it is not the data that is valuable; it is the intelligence and insight it gives us. This insight helps major retailers run their stores and make long-term decisions about suppliers, workforce and their place in the market.

'The healthcare sector is still some way behind others that use information for intelligence'

Turning data into useful knowledge to improve patient care is challenging. In healthcare, a huge amount of data is collected and ‘warehoused’ for years in data repositories and medical-record departments, quietly minding its own business.

This is a missed opportunity because that data may contain the information required to improve care, save lives or improve the lives of healthcare workers.

This lack of use happens for a number of reasons. For example, how data is collected and stored can make using it difficult, particularly for clinicians. Many applications used in healthcare are designed to capture and store data for later review but not necessarily to analyse it. There are also regulatory issues on the use of personal data that are likely to become more complex with the introduction in May of the General Data Protection Regulation (GDPR), the most important change in data privacy regulation in 20 years.

Most organisations keep data in separate applications; for example, a hospital might have one data collection tool for incident reporting, another for recording vital signs and others for pathology, medicines and patient administration. They all work differently and collect data in diverse ways.

Reflecting context

Most importantly, they don’t ‘talk’ to each other, which limits the opportunity to use the data to improve care to those with the technical skills required to download, collate and analyse it to generate useful knowledge. This can take a lot of effort and technical expertise that might be unavailable in organisations.

For data to be useful, it must be high enough quality, and the type of data needs to reflect the context of the questions asked of it. It is not much help if a large supermarket looks at data on avocados when it wants to know where and when baked beans are likely to sell, or plans supplies of ice cream but forgets to look at the weather forecast.

Falling behind 

In England, the Five Year Forward View (NHS England 2014) refers to technology as an enabler, but the healthcare sector is still some way behind others that use information for intelligence. It is still a developing field and nurses, who make up the largest part of the workforce, must influence its development to ensure patient safety.

Nursing in all forms is a massive supplier of care, but it is hard to see this in healthcare organisations’ data. Care is rarely captured and the complexity of care is never so; and, where data on nursing activity is collected, it is often over-simplified.

A couple of years ago, I was asked to work out how many district nurses were needed for a particular area in England. We regularly work out optimum caseloads for groups of specialist nurses and it sounded like an interesting and useful challenge. Optimum caseload calculations require that various scenarios and service models are considered to work out the probability of different events occurring, as this affects workload and workforce numbers. Such stochastic models require a great deal of high-quality data to make them work.

To model a workforce for district nurses, the databases they use were accessed and their work was downloaded. They used two common commercial packages that record work and are used in the community, but the data revealed that the databases were capturing only about 15% of the nurses’ workload. The reason for this was that they were not designed to capture nursing work. They did not include nursing interventions, just broad lists of tasks, and did not capture the multiple interventions needed.

The nurses we looked at did on average six interventions per visit but the system captured only one. Therefore, we had to develop a different way of collecting data that more accurately reflected the situation (Jackson et al 2015).

RCNi digital literacy resource collection

Lack of accuracy 

Lack of accurate data makes showing nursing’s enormous contribution to care challenging. Helping others understand the complexity of care and how nursing affects patient and organisational outcomes such as experience, efficiency and safety is often difficult. If data is inaccurate, staffing models or decisions about safety carry that error. Poor data capture of nursing work means that decision makers make false assumptions about nursing and associated risk factors (Leary et al 2016).

'Involvement in informatics and ensuring that the data captured reflect the real world of nurses are vitally important for the profession; otherwise it risks being invisible'

Making decisions based on poor data is sometimes referred to as ‘garbage in, garbage out’, which originated in the early days of mass computing but remains true today. If the nursing profession allows others to define what it does, it risks losing arguments about its value and risks losing its powerful voice to advocate for patients.

There is so much potential to improve care by collecting nurse- and patient-sensitive data. It can be used to calculate workloads, to provide evidence for the value of nursing and to determine where best to use nursing expertise and how to make care safer by, for example, examining staffing, vital signs and safety data together. Involvement in informatics and ensuring that the data captured reflects the real world of nurses, patients and carers are vitally important for the profession; otherwise it risks being invisible.


References


Alison Leary is professor of healthcare and workforce modelling, London South Bank University

This article is for subscribers only

Jobs