| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Jiska Cohen-Mansfield, Research Institute on Aging, Hebrew Home of Greater Washington, 6121 Montrose Road, Rockville, MD 20852. E-mail: cohen-mansfield{at}hebrew-home.org
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Endoflife care • Nursing home • Quality of life • and Advance directives
Several studies have examined the prevalence of specific advance directives. Table 1 describes some of the advance directives' rates as found in groups of elderly persons. As presented in Table 1, the rates of elderly people with living wills ranged from 11.5% to 31%, those with DPOA ranged from 12% to 66%, and those with DNR ranged from 36.2% to 51.5%. The prevalence rates of advance directives may vary by type of directive, study methodology, type of population, and source of information.
|
The types of documentation available in the facility examined include the section of the Minimum Data Set (MDS; Morris et al., 1991) concerning advance directives, which is both standardized and required by legislation, the narrative notes in the chart itself, and the front cover of the chart, which lists those directives considered currently relevant and applicable to the patient.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Participants
Participants were 122 elderly persons who resided in a 550-bed nursing home. The nursing home was involved in the EVERCARE program, and about half the residents were enrolled in the program. Close to 78% of the participants were female with an average age of 88 years. Of the residents, 69% were widowed, 18% were married, 6% were never married, and 6% were divorced (see Table 2). We rated level of cognitive functioning for all study participants via the MDS-COGS. The study participants had a mean MDS-COGS score (Hartmaier, Sloane, Guess, & Koch, 1994) of 4.97, with the range from 0 to 10, where higher scores indicate greater cognitive impairment.
|
The MDS information followed the MDS advance directives format. We prepared forms for data collection from information inside the chart and from the chart's front cover. The MDS section on advance directives includes items pertaining to the designation of the following directives: (a) "living will," (b) DNR, (c) "do not hospitalize," (d) "feeding restrictions," (e) "medication restrictions," (f) "autopsy request,"(g) "organ donation," and (h) "other treatment restrictions." We collected data from the front cover under the following categories: (a) DNR, (b) do not hospitalize, (c) "no ventilator," (d) "no tube feeding," (e) "no IV," (f) "no antibiotic treatment,"(g) "full code," and (h) other treatment restrictions. The information regarding advance directives from inside the chart was organized in the same fashion as the data set on the front cover.
As part of a larger study on medical decisions in the nursing home (Cohen-Mansfield & Lipson, in press), the researchers asked physicians working in this nursing home and involved in medical decision making for these residents about the following advance directives: (a) do not hospitalize, (b) no ventilator, and (c) no tube feeding. Responses were coded as: (a) "yes," (b) "no," (c) "don't know," and (d) "not applicable." Physicians' interview data were available for 57.4% of the residents. Of those, 53% were the residents' regular physicians, and the rest were physicians on call who had been involved in the medical care of an acute event.
In addition, we extracted demographic information from the chart's MDS demographic sections.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
For 28% of the residents, the item other treatment restrictions from the MDS form was checked. However, the MDS does not specify as to what other means. The most frequent category included in the item "other advance directives" on the front cover and inside the chart sources was "hospitalization with the certain restrictions" (such as "for reversible conditions only," "for comfort care," "if not treatable in this nursing home"). Other directives included "no intubation," "no dialysis," and "no invasive procedures."
Prevalence of advance directives on the basis of chart review was generally highest; prevalance was somewhat lower on the front cover of the chart and was lowest on the MDS. Given that those refer to somewhat different subsamples, we conducted additional analyses to clarify the meaning of these discrepancies.
Rate of Agreement Between Sources
To assess the influence of the information source on the presence of the advance directives, we cross-tabulated data from the three different sources and calculated both agreement rates and Kappa statistics (see Table 4).
|
Rates of Agreement Between Physicians' Reports and Other Sources
The analysis of agreement between the physicians' reports of advance directives and the information from the MDS, front cover, and inside the chart is presented in Table 5 for those cases in which physicians said that the directive was either present or absent (i.e., those cases in which they said it was not applicable or they did not know are excluded from the table). Analysis of the no hospitalization advance directive shows that in seven cases, physicians stated that this directive had been enacted, but it was not indicated in the MDS form. There was also one person for whom the MDS indicated no hospitalization, but the physician said hospitalization was allowed. We also found disagreement between the physicians' reports and the data we obtained from inside the chart and from the front cover. There were three residents for whom the physician reported a no hospitalization directive that was not in a chart source (the inside of the chart was a source for two of the residents; the front cover was a source for one person whose information inside the chart agreed with the physician). For two persons, the chart indicated no hospitalization directives (both cases were stated inside the chart and one of those was also indicated on the front cover), although the physician said that the hospitalization was allowed. Of the 14 persons for whom a physician indicated that he did not know whether there were advance directives concerning hospitalization, 4 had a no hospitalization directive in one source in the chart (either on the front cover or inside the chart), but not in both.
|
Physicians' perceptions indicated moderate agreement with the data from the front cover and inside the chart regarding the no tube feeding advance directive. For eight persons, there was an agreement on the no tube feeding advance directive by all available sources. There were, however, seven cases in which physicians reported such a directive, but it was not reflected in the source (three on the front cover, three inside the chart, and one case with both sources). For 17 residents, the physicians reported not knowing whether they had a directive concerning tube feeding, and of these 17, 12 had a no feeding indication inside the chart or on the front cover or on both.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
The results show a high agreement rate among the written sources of information concerning advance directives with a range of 73.4% to 96.9%, depending on the directive. These rates are highest when base rates of the directives are very low, such as no IV or no antibiotics, which are not present for the vast majority of residents. When base rates for the directive are higher, disagreements between sources are more prevalent, involving up to 27% of the cases.
The reasons for discrepancies may lie in the different functions and procedures pertaining to these source documents. The MDS fulfills a regulatory function and is updated quarterly. It is possible that the lower rates found on the MDS are due to changes in directives that occurred after the last update. Directives on the front cover are used to alert staff members to currently applicable directives. There is indeed a greater concordance between materials on the front cover and inside the chart. However, there are still discrepancies, and those discrepancies go both ways (i.e., some directives are on the front cover but not inside the chart, and vice versa). It is possible that directives that are inside the chart and not on the front are not considered imminently applicable. It is also possible that directives found on the front cover and not in the chart are recently updated by staff members who have not yet been able to add the new information to the notes. A note of caution is that conclusions regarding data from a chart (inside the chart) may depend on the strategy one has used while collecting the information; in other words, "chart review is more difficult than it appears on the surface" (Allison et al., 2000, p. 115). Therefore, when data seems to be missing from inside the chart, that may be attributable to the method of chart review. However, data seem less likely to be missing inside the chart than in other sources, and data missing from other sources (i.e., MDS and front cover of the chart) cannot be explained in this fashion.
An example of differences in directives is found in the chart of Mrs. A. In her MDS, only living will was checked. On the front cover, she had no IV, no antibiotics, and no clysis. All of these directives were also inside the chart, as well as DNR, no hospitalization, no ventilator, and no tube feeding. Staff members noted that directives had recently been changed due to the decline in her situation.
The possibility that the differences between the three written sources can be attributed to their different functions and to updating procedures notwithstanding, the data suggest that, both from a clinical standpoint and from a research point of view, these differences need to be examined. From a clinical point of view, the differences between the documents are highlighted when a resident is hospitalized and a nursing staff member notes that the do not hospitalize order is missing from the front chart. From a research point of view, it is important to either survey all sources or to identify the specific source used and acknowledge the limitation of that source.
The discrepancies between physicians' reports and chart reports include information known to the physician but not indicated in the chart. Such discrepancies are probably partly due to new information about directives that are sometimes changed at the time that the condition of the resident changes. At other times, this may suggest a need for the physician to update the chart more frequently. Conversely, there are times when physicians do not know whether such directives exist, and a substantial proportion of those instances of knowledge deficits involve directives. These directives may or may not be relevant to the decision at hand, but those gaps in knowledge underscore the importance of having the information on the front cover of the chart, where it should be readily available at the time of decision making. The limitation of these data is the small sample size of the physicians' data available for the analysis.
The importance of frequent updating of the chart is highlighted by the fact that close to half (47%) of decision-making processes involving acute events are handled by the physician on call rather than by the resident's regular physician. If the physician who learns about changes in directives does not correct the chart immediately, it is likely that the physician who will next treat the resident will not know about the change in directives. As charts become more computerized, logical links between sources of information can be used to highlight discrepancies and resolve them immediately. The issue of the necessary human input to update those records still requires separate procedures.
Nursing homes may want to use the model described in this article to analyze discrepancies in their charts. Discrepancies that are considered clinically meaningful call for a quality improvement process in which procedures for updating chart information may be revised, or the forms in which such information is captured may be standardized and modified to become more user friendly, more consistent, and most important, more accurate.
|
Footnotes |
|---|
2 George Washington University Medical Center, Washington, DC.
3 Georgetown University Department of Psychology, Washington, DC.
Decision Editor: Laurence G. Branch, PhD
Received for publication November 27, 2001. Accepted for publication June 27, 2002.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. M. Weinick, S. R. Wilcox, E. R. Park, R. T. Griffey, and J. S. Weissman Use of Advance Directives for Nursing Home Residents in the Emergency Department American Journal of Hospice and Palliative Medicine, June 1, 2008; 25(3): 179 - 183. [Abstract] [PDF]
|
|
|
|
 |
|
 J. Cohen-Mansfield and S. Lipson Which Advance Directive Matters? An Analysis of End-of-Life Decisions Made in Nursing Homes Research on Aging, January 1, 2008; 30(1): 74 - 92. [Abstract] [PDF]
|
|
|
|
 |
|
 W. J. McAuley, R. J. Buchanan, S. S. Travis, S. Wang, and M. Kim Recent Trends in Advance Directives at Nursing Home Admission and One Year After Admission. Gerontologist, June 1, 2006; 46(3): 377 - 381. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
