Confused patients present a management problem for nursing staff. The stress level of the nursing staff is raised when increased time and supervision is required to maintain the safety and improve the well-being of confused patients. Also, confusion during hospitalization is associated with increased morbidity, an increased length of stay, and an increased need for nursing home placement after discharge (Inouye, Van Dyck, Alessi, Balkin, Siegal, & Horwitz, 1990), making this a major dilemma for both the health care consumer and health care provider. However, despite the fact that nurses are concerned about the phenomenon of confusion, there have been few nursing studies done to describe the magnitude or prevalence of the problem from the nurse's perspective.
REVIEW OF RELATED NURSING RESEARCH
Nursing research on confused patients has been focused on the assessment and treatment of confusion. The categories of studies include determining nurses' descriptors of confused patients' behaviors; the prevalence of confusion; and descriptions of treatments.
In four studies, nurses' descriptors of confused or agitated patients were identified (Evans, 1987; Ryden, Bossenmaier, & McLachlan, 1991; Struble & Sivertsen, 1987; Yeaw & Abbate, 1993). Investigators found that nurses used combinations of cognitive, behavioral, and verbal terms to describe and identify the patients.
Two research teams measured the prevalence of some component of confusion or agitation. Ryden et al. (1991) determined that 86.3% of the cognitively impaired nursing home patients she studied showed some form of aggressive behavior within a 1-week period. Evans (1987) found that 12.4% of the patients she studied in a nursing home "sundowned." She described sundowning as agitation, restlessness, and confusion that occurs in the evenings. Although 82% of the sundowners were diagnosed with dementia, 85% of the sample diagnosed with dementia did not sundown in her study.
In most of the treatment studies, researchers observed or described the nursing interventions that were used with confused or agitated patients. These interventions included medications and restraints (Morse & McHutchion, 1991; Struble & Sivertsen, 1987), physical characteristics of the facility, staffing patterns, education of the staff, behavioral modification techniques (Herbei, Scherrnerhorn, & Howard, 1990), seizure precautions, one-to-one nursing care, decreasing environmental stimulation, and providing the patients with orientation to their surroundings with both verbal reminders and an orientation chart (Flanigan, 1986).
Although acute care nurses are concerned about confused patients and the difficulties in managing their care, it is unclear whether that is because these patients are numerous, or because each patient presents an overwhelming challenge.
Few studies have evaluated the prevalence of confusion, and even fewer describe the factors affecting confusion in the acute care setting. Therefore, it would be warranted to determine the prevalence of confusion in acute care, as well as to explore the associated assessment and treatment challenges of mese patients.
This study was designed to take advantage of a facility relocation to investigate the effect of environmental factors on the prevalence of confusion. The purposes of the study were to determine nurses' assessment of the prevalence of confused patients at an institution before and after relocation to a new facility, to determine whether environmental factors in the new facility (more windows, private rooms, clocks and calendars, and closer proximity of patients) were related to the prevalence of confused patients, and to describe the characteristics of the confused patients.
The survey employed multiple cross-sectional data points in a descriptive design. The sample included all patients on three medicalsurgical units. A subsample of patients were those patients identified as confused by the nurse providing their care at the time of data collection. The data collection occurred in 1991, prior to facility relocation and in 1992, 5 months after relocation. The collection site was a 420-bed, rural, tertiary-care center.
The instrument used was adapted from the Confusion Assessment Method (CAM) tool (Inouye et al., 1990). This instrument elicited data about the patient's confusion history, current mental status, and thinking patterns. It was adapted by adding questions related to use of medications, sitters, and restraints; documentation practices around confused patients; and patient outcomes. The original instrument, tested at two sites, was found to be reliable and valid (Inouye et al., 1990).
The data were collected at two distinct time periods. During each time period, the data were collected on 7 evenings in a 7-week period. Because the hospital's average length of stay (LOS) was 7.2 days, the data collection points were 8 days apart with the resulting schedule of Monday one week, Tuesday the next week. Wednesday the next week, and so on. The data were collected from three hospital units that were thought to have a high prevalence of confused patients: medical hematology-oncology, neurology, and general surgery. The researcher gathered demographic information on all the units' patients. The researcher then asked each nurse if any of the patients they were caring for that evening were confused. The researcher completed a data collection tool for each identified confused patient. This procedure was followed from March through May 1991 and repeated 5 months after the move to the new facility from March to May 1992.
Demographic Characteristics of the Sample
Data (at least demographic) were collected on both confused and nonconfused patients for a total of 1,200 patient records during this time, with more extensive data collection of the confused patients. Because several of the patients were captured in more than one data-collection time, repeated cases were removed.
There were 843 patients in the final sample; 435 patients (52%) in 1991 and 408 (48%) in 1992. Demographic characteristics for the two years are summarized in Table 1. Patients ranged in age from 10 to 92 with a mean age of 59. There were slightly more women (53%) than men (47%) and the average length of stay was 16 days.
Nurses' assessment of the prevalence of confusion
The prevalence of confusion according to nurses' judgments ranged from 3.9% to 13.7% for the 14 data collection times. In the 1991 data collection period, 45 patients (10.3%) were identified by nurses as confused. In the 1992 sample, 37 patients (9.1%) were identified as confused. In the 2 years combined, 761 patients (90.3%) were not confused and 82 patients (9.7%) were confused. The neurology unit had the highest prevalence of confusion (15.7%). This rate was more than twice that of either medical hematology-oncology (7.5%) or general surgery (7.3%) units. After the duplicate cases were removed, the final sample of confused patients (n = 82) contained 37 (45%) from neurology, 22 (26%) from medical hematology-oncology, and 23 (28%) from general surgery. Based on chi-square analysis, these differences by unit were significant (Xp 2 = 13.2, rf/=2, p = .001).
The new facility was designed with more private rooms than the old facility. More patients in the new facility were in private rooms (69%) than in the old facility (50%). Fewer patients were housed in four-bed wards in the new faculty (3% vs. 8%). The mean number of roommates of confused patients was 0.7; for nonconfused patients, 0.6, not significantly different. Most confused patients in this study had been on the same unit (91%) and in the same room during the previous week (92%). Chi-square analysis revealed no significant difference in the rate of confusion between 1991 and 1992 (Xp 2 = 39, df=l, p = .53). Because this difference was not significant, the remaining results will be reported for the sample as a whole, combining 1991 and 1992.
Length of Sf ay* by Age Group
Characteristics of confused patients
When compared to the group of non-confused patients, confused patients were significantly older (i = -3.05, ? = .002). The mean age for confused patients was 64.5 years (SD = IS) and for non-confused patients the mean was 58 (SD = 17).
Confused patients also had a significantly greater LOS than did the group of non-confused patients (i = -4.26, ? = .001). The mean LOS was 30.7 days for confused patients (range = 1 to 203 days; SD = 33.6) and 14.4 days for non-confused patients (range =! to 432 days; SD = 24). The mean LOS for confused and nonconfused patients is summarized by age groups in Table 2. For all age groups, confused patients had a longer LOS than non-confused patiente.
History of confusion There were 66 confused patients whose mental status history was known: 35 patiente (53%) were categorized as acutely confused, that is typical mental status at home, but with acute changes during hospitalization. The rest were categorized as chronically confused, with or without changes during this hospitalization. The mean age for the acutely confused patients, 58 (SD = 20.1), was significantly younger than that of the chronically confused patients, 69.5, (SD =16; f = -2.55, p = .013). The mean LOS for chronically confused patiente, 33.2 days, (SD = 35) was longer than that of the acutely confused patients, 23.3 days, (SD = 16), but the difference was not statistically significant. When an outlier with a LOS of 153 days was excluded from the chronically confused patiente, the difference remained non-significant.
Restlessness Restlessness was ascertained by asking the question: "Does the confused patient have an unusually high level of motor activity during this shift?" Thirty-six patients (48%) were identified by the nurses to have this characteristic. The interventions that were used with the patient with recent restlessness included restraints for 19 (53%), sitters for 8 (22%), and medications for 17 (47%) of the patiente.
Information was collected on three of the common interventions used with patients who are confused: medications, restraints, and use of sitters. Of the three, sitters were used much less frequently (17%) than either restraints (38%) or medications (38%).
Use of medications Thirty-one confused patients had orders for medications to treat agitation. Haldol was the most frequently ordered medication (n = 12). Of those who received the drug within the previous 24 hours, the mean total dosage re'ceived was 19.5 mg (range = 5 mg to 50 mg). Eight (26%) received Ativan with a mean total dose of 3.4 mg (range= .5 mg to 16 mg). Serax was ordered for two patients (6.5%), Librium for two (6.5%) and "other" med' ications for six patients.
More chronically confused patients (45%) had medications ordered than did those who were acutely confused (32%), but the difference was not significant.
Use of restraints Restraints were used on 31 (38%) of the confused patients during the current shift. Of the various types, the two most commonly used were the chest or posey restraints on 13 and the two-point restraint on 10 patients. Additionally, five patients also had a second type of restraint (chest or posey restraint) used during that shift. Of the 31 patients who were restrained, eight patients also had sitters and of those, five patients had medications as well. A total of 11 of the patients who were restrained received medications. There were no significant relationships between restraints or medications and sitter use. The use of restraints in the acute and chroni. cally confused patients was 40% and 42%, respectively, which was not significantly different.
Restlessness was a common characteristic in patients, who had recently been restrained. Of the patients who were restrained, 19 (61%) were identified as being restless during the current shift. Patients with a high level of restlessness were significantly more likely to be restrained than patients who were not restrained (Xp 2 = 6.89, df= l,p = .009).
Patient Outcomes af Time of Discharge
Use of sitters Only 14 (17%) of confused patients had some type of sitter in the previous 24-hour period. Hired sitters were used for 10 patients and family members functioned as sitters for four patients. Six (17%) of the acutely confused and six (19%) of the chronically confused used some type of sitter. The difference in use of sitters between acute and chronically confused was not significant. In addition, no relationships were found between restraints, or restlessness and the use of sitters. The only intervention found to be significantly related to sitter use was that of the use of medications (Xp 2 = 5.19, d/=l, p = .022). Patients who used medications were more likely to use sitters than those who did not.
Patient outcomes at discharge
Patient outcomes at discharge were coded into four categories:
* home with a visiting nurse referral,
* transfer to another institution, or
Patient outcomes were not significantly different between 1991 and 1992. Confused patients had less favorable outcomes at discharge than did the group of non-confused patients. Outcomes for the confused and non-confused patients are summarized in Table 3.
Confesion and patient outcomes Of the non-confused patients, 60% went home, 24% went home with a visiting nurse referral, 10% were institutionalized, and 4% died. Of the confused patients, fewer went home (9%); more went home with a visiting nurse referral (34%), were institutionalized (44%), or died (11%). Based on chi-square analysis, there was a significant relationship between confusion status and outcome at discharge (X2 = 113.89, d/=3, p<.0001). In general, patients who had been acutely confused in the hospital seemed to have better outcomes than those who had a history of chronic confusion. When compared to the chronically confused patients, more of the acutely confused patients went home (9% vs. 7%) or went home with a visiting nurse referral (46% vs. 20%). Also, fewer acutely confused patients went to another institution (40% vs. 57%) or died (6% vs. 17%) then chronically confused patients.
Documentation of Confusion
Evidence of recent documentation related to confusion was found in the nurses' notes of 63 (78%) of the confused patients. Only 34 (42%) of the patients identified as confused had documentation of their confusion status in their nursing care plan. None of the interventions used with confused patients - restraints, sitters, or medications - was related to recent documentation in nurses' notes or nursing care plan. Restless patients had significantly greater rates of documentation in nurses' notes than did non-restless patients (X2 = 4.37, df= 1, p = .002), but this was not true of documentation in the nursing care plan.
The longer LOS and poorer outcomes at discharge for the confused patients in this study could be due to several factors such as age, coexisting medical conditions, or the confusion itself. Because the confused sample of patients was older than the non-confused sample, age might have been a contributing factor to the outcomes. Regardless of the cause, the health care costs for these older confused patients were much higher due to their extended length of hospital stay and degree of assistance required post-hospitalization than for the younger non-confused patients. Was the greater LOS related to the patients' being a placement problem, perhaps due to their confusion, or was it due to medical complications requiring longer hospitalization, potentially from the same etiologies as their confusion? Teasing out these causal relationships would help to guide appropriate strategies for intervention, and thus help guide the allocation of scarce resources more wisely.
The question of whether environmental factors, such as the design of the facility, would affect the prevalence of confusion remains a question because the design of this study relied on the staff nurses' assessment of whether the patient was confused. The layout of the old facility might have hampered the nurses' awareness of the patient's confusion due to lack of visibility whereas the new facility's layout might have heightened the ability of the nurses to observe and become aware of confusion in their patients. Also, the increased potential of the nurses in the new facility to interact with patients, due to a closer arrangement of the nursing stations and the patient rooms, may have allowed the nurses in 1992 to take notice of those patients who were confused, but did^ not have obvious behavioral signs (these patients were classified as the quietly confused). These could have lowered the 1991 numbers artificially and raised the 1992 numbers to the more accurate level, thus masking' any significant difference in the facilities. Conversely, the nurses' observational abilities might not have been altered by the design changes, and therefore it could be concluded that the design of the units might not affect a patient's potential for becoming confused. It may be that environmental factors do not affect the prevalence of confusion per se, but rather the outcomes for these patients.
Nursing interventions used with these confused patients included medications, physical restraints, and sitters. Haldol was the most commonly used medication. The mean total daily dose (19.5 mg) was high for an elderly population. However, these - data were not analyzed by age group and included younger patients.
Restraints were used with as many as 38% of the confused patients. The issue of restraint use has come under question recently, first in long-term care and now in acute care. A Food and Drug Administration Safety Alert (USFDA, 1992) pointed out increasing national concerns related to potential hazards to patients subjected to restraints. More long-term « care facilities are embracing a "restraint-free" environment policy. Our study showed a relationship between highly restless patient behavior and restraint use. We must ask the question: Were the restraints used to control this behavior, or were they causing or increasing the behavior?
Finally, in an era of cost containment, are interventions chosen for efficacy or cost savings? The choice of restraint over sitter use might reflect the lower immediate cost of the restraint over an extra staff member who sits with the patient; however, these cost savings may only reflect immediate costs and not take into account injuries or longer states of confusion due to restraints. Acute care facilities struggle with the issue of restraint use, especially because their patients often have tubes and drains that could become dislodged.
Our study did not show a statistically significant difference in the care of chronically confused patients versus acutely confused patients; however, the chronically confused patients tended to have poorer posthospitalization outcomes than did the acutely confused patients. This makes intuitive sense because one would hope that the acutely confused patients should recover from their confusion prior to discharge. These results should be viewed with caution, however, because our study design defined patients as having chronic confusion even if their confusion started just prior to hospitalization. This definition might have artificially increased the sample of patients classified as chronically confused, and decreased the sample of those classified as acutely confused.
The findings related to the extended LOS for confused patients have clear implications for nursing care. Regardless of the factors influencing the LOS for confused patients, nurses in acute care settings must consider alternate care arrangements to meet the needs of this population consistently and efficiently. Perhaps a particular area or wing of the institution could be used as an intermediate placement area for patients with delayed discharge due to disposition complications. This particular area of the hospital could incorporate environmental and long-term care strategies for dealing with confusion in the provision of care.
This study highlights the need for continuing nursing education related to confusion in the acute care setting. The prevalence of confused patients in this study was 10%, yet it seemed difficult for nurses to identify some of the confused patients. In addition, the nurse often was uncertain about when the patient first became confused. The patient may have experienced the acute onset of confusion before hospital admission due to an illness process. An accurate confusion history is essential in order to gauge the patient's improvement or the patient's return to baseline state.
The findings of the study related to interventions mat are used for confused patients and documentation issues around confusion also indicate educational issues. Educating nurses in the acute setting can be directed by the nursing process. In this way, an educational program would consist of teaching confusion assessment skills with the intent of increasing nurses' awareness of confusion in their patients, as well as increasing nurses' awareness of the importance of the patient's confusion history. Additional teaching components would include instruction about interventions for confusion, as well as methods to evaluate how well these interventions might be working. This total educational approach would be reflected by more appropriate identification and treatment of confusion in the acute setting.
It may be that part of the uncertainty about the identification of confusion in this study is related to the current lack of consensus about definitions of and treatments for confusion in the acute setting, as well as problems with the documentation of such information. The nurses often knew the patient was confused, but documentation about this point was inconsistent. This may be due in part to documentation systems that are cumbersome, such as the nursing care plan. A potential solution to some of these problems may be found when computerization of nursing documentation becomes more common.
This study points out the importance of initiating discharge planning from the time of admission. Confused patients are significantly more likely to use visiting nurse services or to need placement at discharge. Placement can be problematic for this population and alternate methods of care at discharge should be investigated and encouraged if feasible. The aggressive use of home health care, in conjunction with the use of paid caregivers, might be more economical for the mildly confused, stable, or improving patient than placement in an extended-care facility.
In light of this, it seems important for nurses to become involved in affecting social and legislative policy decisions that determine the allocation of resources for post-discharge care of confused patients. Also, additional research should focus on alternate care arrangements or programs intended to increase the skills of nurses in acute care settings for providing care to confused patients.
In considering the design of this study, two limitations became apparent. The tool used for this study, the CAM (Inouye et al, 1990), was intended for use in identifying delirium. The use of this tool in a chronically confused sample may alter the validity of the tool. A second limitation is that the researchers did not actually assess the patients identified as confused. An assumption was made that the nurse's assessment of confusion was accurate. This may not, in fact, have been the case and this would then alter the prevalence data.
The average age of patients in acute care settings will probably increase as the population continues to age. In this study, confused patients were significantly older than those who were not confused. Based on these results and the trend of the population to include greater numbers of elderly, it is probable that we will see an increase in the overall percentage of patients in acute care who are confused. Therefore, it is safe to assume that confusion will remain a significant issue in our acute care settings. We must continue to seek ways to meet the nursing care challenges posed by the needs of these vulnerable patients.
- Evans, L.K. (1987). Sundown syndrome in institutionalized elderly. Journal of American Geriatrics Society, 35, 101-108.
- Flanigan, M.D. (1986). Clinical management of pediatrìe agitated dosed head injured clients. Unpublished master's thesis. University of Washington, Seattle, WA.
- Herbei, K., Schermerhom, L-, & Howard, \ (1990). Management of agitated headinjured patients: A survey of current techniques. Rehabilitation Nursing, 15(2), 66-69.
- Inouye, S.K., Van Dyck, C.H., AJessi, C-A., Balkin, S-, Siegal, A.P., & Horwitz, R.I. (1990). Clarifying confusion: The confusion assessment method. Annals of Internal Medicine, 333,941-948.
- Morse, J.M., & McHutchion, E. (1991). Releasing restraints: Providing safe care for the elderly. Research in Nursing and Health, U1 187-1%.
- Ryden, M.B., Bossenmaier, M., & McLachlan, C. (1991). Aggressive behavior in cognitiveiy impaired nursing home residents. Research in Nursing and Health, 14(2), 87-95.
- Struble, L.M., & Sivertsen, L. (1987). Agitation behaviors in confused elderly patients, journal of Gerontobgical Nursing, 33(11), 40-44.
- U.S. Food and Drug Administration. (1992). FDA Safety Alert: Potential hazards with restraint devices. U.S. Department of Health and Human Services, Rockville, MD.
- Yeaw, E.M.J., & Abbate, J.H. (1993). Identification of confusion among the elderly in an acute care setting. Clinical Nurse Specialist, 7, 192-197.
Demographic Characteristics of the Sample
Length of Sf ay* by Age Group
Patient Outcomes af Time of Discharge