Haematologist/ Honorary Senior Lecturer
Newcastle Hospitals Trust
Newcastle upon Tyne, UK
Mrs Joan Smith is in hospital awaiting an operation. In the next bed Mrs Jean Smith is recovering from her operation. A nurse is sent to take blood from Mrs Joan Smith to test the blood group and prepare some blood that may be necessary during the operation. She goes to the wrong patient and asks carefully, “Are you Mrs Smith, Mrs Joan Smith?” The patient, Mrs Jean Smith, drowsy with pain relief, replies “Yes” and the blood is taken but incorrectly labelled for the real Mrs Joan Smith. Blood of the wrong group is prepared and during the operation is given to the patient. The result is fatal. This fictional scenario will be alarmingly realistic to many doctors and nurses.
In modern hospitals patients move rapidly from one department to another for blood tests, X-rays, endoscopies, and so on. They may receive prescriptions or intravenous drugs or be subject to therapeutic procedures. At each stage a doctor, nurse or technician may meet them for a short time only, so correct identification of a patient becomes vital. Although many patients can be relied on to identify themselves accurately if asked, some are old or hard of hearing, some confused and some unconscious. Many are scared and in unfamiliar surroundings. A test carried out on the wrong patient may have serious consequences, perhaps for patient A on whom it was done in error, or for the intended patient B in whose notes the result from patient A is filed.
Blood transfusion is a good example of a procedure where a single error of identification can have fatal consequences. An error in patient identification when the sample for determining blood group is taken, in booking-in that sample in the laboratory or in administering a unit of blood can result in a serious or life-threatening reaction. Each year we record several sampling errors in our hospital, despite strict protocols and repeated education. Each year a confidential reporting scheme for transfusion errors in the UK (SHOT) records many serious and a few fatal reactions due to such errors in patient identification, either at the point of sampling or at the point of administration.
Introducing complex checking procedures to prevent such errors may seem a good idea, but paradoxically it can have the opposite effect as staff take shortcuts or fail to concentrate on the important points in an overcomplex protocol. Any checking process should be simple to learn and follow and should concentrate on the important points.
In the UK it is standard practice for all patients in hospital to have a small paper band around their wrist on which are recorded the salient points of identity: first name, surname, gender, date of birth, a unique hospital number not shared with any other patient, and religion. This wristband acts as the critical identifier of the patient. In 2002 we introduced a simplified system for checking blood administration that we called the “Tag and Label” system. The local success of this prompted us to design a small machine that could both confirm the checks electronically and improve the written record of the check. This requires a machine-readable code attached to the patient. At present the simplest and most robust method for this is a wristband on which the unique hospital number is given in both eye-readable and barcode format. Barcodes proliferate in medical notes and are merely a robust method of recording a number that can be read by a machine. Simple reading of a barcode does not confirm identity if that barcode is attached only to a piece of paper. Therefore, we asked that the barcode on the patient’s wristband should have a prefix that identified it as being on a wristband. The electronic checker would accept that barcode only if it registered the correct prefix first – that would not prevent malicious attempts to misidentify a patient, but it would prevent casual failure due to sloppiness.
Systems already exist for using barcoded wristbands to label blood samples taken from a patient, but the use of “barcodes with identity” will further improve safety by ensuring as far as possible that the operator is reading directly from the patient when he or she labels blood samples, administers transfusions, gives chemotherapy or authorises an X-ray.
The device for checking blood at the bedside using barcoded wristbands and barcoded blood bag labels was tried out in our daycase unit for patients with blood disorders in our own hospital and in a neighbouring hospital. At both sites the nurses preferred using the barcode checker rather than the standard manual checking protocol. Reasons given were that it was quicker, they could do it with confidence on their own (most hospitals use two nurses to cross-check each other) and it clearly recorded the fact that the check had been done and proved correct on a printed label that could then go in the patient’s case-notes as a record of that transaction. The result should be reduced errors, a better printed record and reduced nursing time leading to cost saving. Our experience leads us to believe that use of electronic patient identification will become standard for many patient transactions in the near future.
Photo ID cards, and beyond
In an interesting development of this concept, we have been part of a trial using photo ID cards for patients who regularly attend hospital, such as patients requiring regular blood transfusion. This idea originated at the Central Middlesex Hospital in London. Using easily available software linked to the hospital patient database, a small plastic card is printed containing details of the patient – including a small photo taken with a camera linked to a PC. This card can replace the wristband. The advantages are a photographic means of identification, empowering the patient and its reusability. As on the wristband, the unique hospital number can also be given in barcode format. The trial was run by the UK’s National Patient Safety Agency and found the photo ID cards to be very acceptable and popular with patients. We have recently introduced them as a standard method of patient identification in our haematology daycase unit, where many patients present for outpatient blood transfusion, chemotherapy and other procedures. We expect that the idea will become widespread in hospital departments, such as chemotherapy units or kidney dialysis units, where patients attend repeatedly.
Linear barcodes are only one form of machine-readable script. 2D barcodes, portable data files and radio-frequency identity devices can all perform a similar function.
We think it is likely that wristbands or identity cards, including machine-readable details, will become standard practice in many hospitals over the next few years. Small devices for checking all forms or transactions between patients and hospital staff will become widespread. The result should be fewer errors and, it is hoped, fewer serious consequences.