​
​Photo: Mikkel Thor Olsen is one of 26 experts from Denmark and abroad behind the first consensus report on hospitals’ use of diabetes technology for inpatients. Photograph: Lizette Skotbo Kabré.​

Patients with diabetes should normally be able to keep their glucose sensor or insulin pump when they are admitted to hospital. For some patients, hospitals should even consider introducing diabetes technology during their stay, although they were not using it before admission.

These are among the main messages of a new international consensus developed by 26 ex-perts from eight countries. The consensus article has been published in the two respected journals Diabetologia​​ ​and Diabetes Care.​

The consensus group has drawn up recommendations on how hospitals can use continuous glucose monitoring, insulin pumps and automated insulin delivery systems in adults with diabetes. The recommendations cover, among other things, which patients can continue using their own equipment, who is responsible for it, how staff should respond to sensor readings and alarms, and when the technology should not be used.

One of the experts behind the consensus is Mikkel Thor Olsen, a doctor and researcher at Steno Diabetes Center Copenhagen. His research focuses on the use of diabetes technology during hospital stays.

“One of our most important messages is that normally patients do not need to have their diabetes technology taken away simply because they are admitted to hospital. But hospitals need clear procedures for who can use the technology, who is responsible, and how staff should respond to the readings,” says Mikkel Thor Olsen.

Close monitoring replaced by three finger-prick tests a day

The recommendations come at a time when diabetes technology has significantly changed treatment outside hospitals. A glucose sensor can measure glucose around the clock, providing a new reading every few minutes. Automated insulin delivery systems can use these readings to adjust insulin delivery continuously.

Around one in five hospital inpatients have diabetes, and blood glucose levels in hospital are still often assessed using three to five finger-prick tests over a 24-hour period. This means that patients can go from very close monitoring at home to far fewer measurements at precisely the point when they become ill and are admitted to hospital.

“The technology has developed much faster than hospital procedures for using it. Virtually all diabetes technology has been developed for use outside hospital, so simply bringing it into a hospital room is not enough. We also need to build the system around it,” says Mikkel Thor Olsen.

He points out that this is particularly relevant because blood glucose often becomes harder to manage during a hospital stay. Acute illness can cause blood glucose to rise, while reduced physical activity, changes to meals, fasting before surgery, tube feeding and treatment with medicines such as corticosteroids can alter insulin requirements from hour to hour.

According to Mikkel Thor Olsen, high, low and widely fluctuating blood glucose levels are all known to be associated with poorer outcomes during hospital stays, including more complications and longer stays.

Use should not depend on the individual department 

People who are used to monitoring their blood glucose continuously with a sensor at home may find that, during a hospital stay, staff still base treatment on occasional finger-prick tests. An insulin pump or automated insulin delivery system that a patient normally manages independently can suddenly raise questions about responsibility, safety and procedures.

Can the sensor stay in place? Who monitors the alarms? And who adjusts treatment if blood glucose rises or falls rapidly? In practice, the answers can vary from department to depart-ment and depend on how much experience individual staff members have with the technology. This exact variation is what the consensus group wants to reduce.

“We want to move away from an approach where the way the technology is managed de-pends on the individual department, staff member or patient. If it is to be used safely, there needs to be a clear plan covering the entire hospital stay,” says Mikkel Thor Olsen.

New patient groups should be offered sensors during hospital stays

The new consensus is not only about people who already use diabetes technology. The ex-perts also identify groups for whom hospitals should consider introducing continuous glucose monitoring during their stay.

These may include patients treated with insulin and people at particular risk of low or high blood glucose. Older patients, people with kidney disease or several coexisting medical conditions, and patients who have previously experienced severe episodes of low blood glucose may be among those who benefit particularly from closer monitoring. The same applies to patients receiving treatments that make blood glucose difficult to manage, such as cortico-steroids or tube feeding.

Sensors could therefore move beyond being primarily patients’ own tools to become an active part of hospital treatment as well.

“A sensor provides far more information than conventional measurements and can make it possible to detect problems earlier. But the wealth of data is only useful if staff know how to respond to it,” says Mikkel Thor Olsen.

Technology should never be used as a matter of principle

The consensus group also emphasises that technology should not be used uncritically. If a patient already uses a sensor, insulin pump or automated insulin delivery, this should generally continue, provided that the patient or staff can manage the equipment safely. However, that assessment must be open to revision during the hospital stay. Patients who can manage their own equipment on admission may become more seriously ill and later lose the ability to do so.

There are also situations in which use of the technology should not continue. These include diabetic ketoacidosis and similar acute conditions, situations where neither the patient nor staff can manage the equipment safely, and certain diagnostic imaging procedures. According to Mikkel Thor Olsen, the guiding principle should not be that technology must always be used, but that decisions about its use should be made systematically and on sound clinical grounds.

Different blood glucose targets apply in hospital

An important point for the expert group is that hospitals should not aim to control blood glucose as tightly as in patients’ everyday lives. In long-term diabetes care, good blood glucose management is partly about reducing the risk of complications over many years.

“Blood glucose can change considerably during a hospital stay because of acute illness, surgery, changes in food intake and new medication. We should therefore not necessarily pursue the same exact targets as at home. In hospital, the main short-term priority is to avoid low and very high blood glucose levels,” says Mikkel Thor Olsen.

The consensus group proposes specific targets for sensor-based glucose monitoring during hospital stays. For most patients, glucose levels should be between 3.9 and 10.0 mmol/L for more than 60% of the time. They should be above 10.0 mmol/L for less than 25% of the time and above 13.9 mmol/L for less than 5%.

Limiting low glucose levels is particularly important. Time below 3.9 mmol/L should be zero wherever possible. The group also recommends that glucose levels should be between 3.9 and 5.6 mmol/L for less than 15% of the time, as readings at the lower end of the normal range can be an early warning that glucose is heading too low. However, it must always be possible to adapt the targets to the individual patient.

Technology requires new ways of working

The biggest challenge for hospitals is to establish a structure in which sensors or insulin pumps become a natural part of everyday care.

The consensus group recommends, among other things, standardised procedures covering which patients can continue using their own equipment, how sensor alarms should be man-aged, and when readings should be checked using conventional blood glucose measurements.

Staff training is also needed, along with greater integration of sensor data into electronic patient records.

“Hospitals cannot simply take the technology patients use at home and expect it to work automatically within a hospital organisation. They need to build a structure around it – with clear responsibilities, training and shared guidelines,” says Mikkel Thor Olsen.

For Mikkel Thor Olsen, the implications therefore go beyond whether a sensor can remain on a patient’s arm during a hospital stay. It is about hospitals adapting to diabetes care that has already changed significantly beyond their walls.

“The aim is for the use of diabetes technology not to depend on which department a patient happens to be admitted to, or whether an individual staff member is familiar with the equip-ment. It needs to become a more systematic and standardised part of care throughout the hospital stay,” says Mikkel Thor Olsen.