Fræðigrein

Læknablaðið 2005; 91: 153-60

Background: A municipal nursing home with 68 beds in Reykjavík, opened in mid-year 1982.

Objectives: To analyse changes in demographic, health and outcome variables over 20 years.

Design: Retrospective analyses of data from medical records of all diseased persons with cross-sectional comparison of five four-year intervals.

Setting: Droplaugarstadir Nursing Home in Reykjavík. The nursing home is supervised by fully qualified nurses and provides maintenance rehabilitation. Medical services are delivered from a specialist geriatric hospital department.

Residents: All residents who died 1983 to 2003.

Measurements: Demographic data, type of dwelling before admission, Nursing Home Pre-admission Assessment Score (NAPA) (*), mobility- and cognitive score, drug usage and a list of medical diagnoses. All recorded health events during stay, falls and fractures, medical and specialist consultations. Advance directives, as recorded and end-of-life treatment, place of death, clinical diag­nosis of cause of death and length of stay.

Results: The total number of medical records read numbered 385, including 279 females and 106 males. The mean age on admission was 85 (±7) years. During the first 4 years the majority of residents came from their own private homes or residential settings but in the last four years, 60% were admitted directly from a hospital ward. The mortality rate was 17% per year in the first period and the majority died in a hospital. This ratio took a sharp turn as the mortality rate increased to 40%, and in the last period only 2 of 97 deaths took place in a hospital. Admission mobility- and cognitive scores showed increased disability with time. The most common diagnosis on admission was dementia (56%), ischemic heart disease (46%), fractures (35%) and strokes (27%). Parkinsonism and maturity onset diabetes had a low prevalence rate of 6%. A mean NHPA of 57 (±17) points confirmed a high dependency selection. The mean number of drugs per patient was 5.3 (±3), including 1.1 (±1) for psychoactive drugs and sedatives. The most common health events during residents? stay were urinary and respiratory infections, heart failure, cardiac- and cerebral events and pulmonary disorders. Hip fractures occurred in 45 residents (12%) and other types of fractures in 47 during their stay in the nursing home. The number of medical visits and specialist referrals increased with time. Palliative care was the most common form of treatment at end of life. Pneumonia was most commonly recorded cause of death in medical notes. The yearly mortality rate was 29% and the mean length of stay was 3 (± 2,9) years for the whole period.

Limitations: Retrospective analyses have many inherent drawbacks and the information in medical records tend to be scanty. Analyses of disabilities, as described in the medical record, can only be descriptive and health events are likely to be underreported. Statistical methods have a less meaningful role for interpretation as only diseased persons were included and survivors excluded. However, the length of time, uniform medical care and turnover rate of residents generate useful information on the patterns of the nursing home service during a time of considerable change.

Conclusions: This retrospective analysis indicates increasing frailty in nursing home patients admitted over a period of 20 years. With time the residents are more often admitted directly from a hospital rather than from an individual dwelling. Most deaths took place in the nursing home and were preceded with informal or formal palliative care directives, which was a significant change over time. The data indicates growing efficiency in the nursing home selection processes due to the NHPA and improvements in holistic geriatric care. This development is in keeping with the Icelandic health care policy for elderly people to stay longer in their own home with access to a nursing home placement when needed.

* Jóhannesdóttir GB, Jónsson PV. Nursing Home Preadmission Assessment in Reykjavik 1992. Artic Medical Research 1994; 53: 512-4.