There were 9 studies of assessment; the design of the studies was unclear. There were 3 studies of management: 2 RCTs and one uncontrolled study.
Assessment of fluid intake (3 studies).
The reviewers looked at these studies from the point of view of how much fluid was taken in, rather than at how effective different measures of fluid intake were. The results are therefore not presented here.
Assessment of dehydration (6 studies).
Omnibus Budget Reconciliation Act Minimum Data Set (1 study): no mention of the reliability or test performance of this tool was provided.
Axillary moisture (1 study, n=86): the axilla was graded as either dry or moist and was compared to a biochemical assessment of dehydration, which was considered the 'gold' standard. The reliability was 0.5, sensitivity 50%, and specificity 82%.
Intra-ocular pressure (1 study, n=13): the correlation of intra-ocular pressure with measures of dehydration was measured. There was no correlation between intra-ocular pressure and changes in serum osmolality or urea concentration.
Febrile episodes (1 study, n=130): the correlation between febrile episodes and biochemical measures was investigated. There was a significant association between febrile episodes and dehydration. However, sensitivity and specificity were not calculated, although the data to calculate these was available.
Physician and nursing evaluation (1 study, n=54): a physician's rating of dehydration was compared with biochemical evidence of dehydration. The physician assessment did not miss any dehydrated patients but did overestimate the number of patients with dehydration. Clinical indicators of dehydration, as assessed in the bedside hydration examination (e.g. longitudinal tongue furrows, sunken eyes, dry mucous membranes, upper body muscle weakness, speech difficulty and confusion), had the strongest correlations with dehydration severity, as determined by biochemical markers.
Biochemical tests (1 study, n=230): urine specific gravity was used as the 'gold' standard. The study found that the serum urea nitrogen-to-creatinine ratio correlated with urine specific gravity as an accurate indicator of early hypo-hydration. Haematocrit and serum osmolality showed no correlation.
Management (3 studies).
Body position of the feeder (1 RCT, n=39): the position of the feeder had no effect on the food and fluid intake levels of residents. Improving hydration (1 RCT and 1 uncontrolled study, n=29): one RCT found that offering fluids 1.5 hourly, throughout the waking day, to bedridden residents maintained hydration at a significantly higher level than did 3-hourly bed checks with no prompt to residents for fluids. One uncontrolled study found that the use of oral hydration solution significantly improved the hydration of older adults.