Better outcomes occur in residents of long-term care facilities if acute illnesses are detected and treated early. For a variety of reasons, residents may not report (or be able to report) symptoms or changes in symptoms that may indicate acute illness. As the primary caregivers in the facility, the nursing assistants are uniquely situated to detect changes that may indicate the onset of early illness. Boockvar and colleagues conducted this observational cohort study to develop a standardized instrument (see accompanying figure) for use by nursing assistants to help in documenting behavioral and functional status changes in nursing home residents and, in turn, alert the medical staff to early signs of acute illness.

The items for the instrument were produced after focus groups of physicians, nurses, nursing assistants and therapists discussed what changes indicated to them that a resident is ill or becoming ill. These answers were then categorized and included in the instrument, which also included patient-specific items, such as how the patient spent time between meals and how the patient moved around. The categories of change were behavioral, functional and cognitive. All items were answered based on current and previous observations. The nursing assistants completed the form for each patient in their care on a daily basis but were also encouraged to continue reporting changes to the nurse or other staff as they usually did. Rounds with a researcher occurred weekly to determine the occurrence of acute illness in each patient.

About two thirds of the 74 patients included in the study had dementia. During the study period, 19 patients (26 percent) developed an acute illness. There were 1,734 evaluations submitted by the nursing assistants. The assessment instrument had a positive predictive value of 17 percent and a negative predictive value of 96 percent for the occurrence of acute illness within seven days. In 80 percent of the acute illnesses that were detected, the nursing assistants documented a status change on the instrument a median of five days before any record of illness was documented in the medical chart. The assessment instrument took no more than 15 minutes per day to complete. The only acute illnesses not detected were acute gastroenteritis, gastrostomy infection and seizure. The assessment instrument detected changes in patients who were later determined to have urinary tract infection, pneumonia/bronchitis, congestive heart failure, upper respiratory infection, liver failure, sick sinus syndrome, gallstone pancreatitis, transient ischemic attack and dehydration. The five items that were included in the final assessment instrument and that were the most predictive of an acute change were weakness, agitation (or nervousness), a self-reported complaint and whether the resident ate the same amount of food and said hello or smiled at the nursing assistants (the latter being a patient-specific item). If the instrument was positive on one or more items (showing a change), there was a fourfold increase in the rate of acute illness.

The authors conclude that the assessment instrument is reliable for detecting early signs of acute illness.