Documentation and Management of Words Associated With Delirium Among Elderly Patients in Postacute Care: A Pilot Investigation

Article Outline

• Abstract
• Methods
• Results
• Discussion
• Acknowledgments
• References
• Copyright

Objective

To describe in a pilot investigation the frequency that keywords associated with delirium were documented by providers and to study the effect of reporting such observations on physician orders.

Design

Retrospective investigation.

Settings and Participants

Eighty elderly patients identified from 895 admitted to 2 postacute care (PAC) facilities.

Measurements

Keywords associated with delirium were confusion, disorientation, altered mental status, delirium, agitation, inappropriate behavior, mental status change, inattention, hallucination, and lethargy. The source of the words and actions taken were recorded.

Results

Keywords associated with delirium were identified in 80 (9%) of 883 patients who met inclusion criteria, with the term “confusion” most frequently noted (95%). Nurses and physicians recorded keywords in 79 (99%) and 55 (69%) patient charts. The actual term “delirium” was used in only 6 (7%) of 80 cases. In 55 (69%) cases when physicians were notified, treatments or evaluations were performed: pharmacological 55 (100%), nonpharmacological 11 (20%), assessments 38 (69%), transfer to the emergency department 19 (34%). Nurses did not alert physicians in 25 (31%) cases where keywords were found and thus no action was taken in these cases.

Conclusions

In this pilot investigation in the postacute setting, nurses and physicians documented words associated with delirium in 9% of the patient charts. When nurses notified physicians of patients with charted keywords suggesting delirium, physicians responded with orders for further assessments or pharmacological interventions. However, nurses did not refer patients with keywords in 1 of 3 cases and no actions were documented in the charts for these patients.

Keywords: Altered mental status, delirium, treatment, postacute, rehabilitation

An increasing number of patients, especially elderly patients, is admitted every year to postacute care (PAC) facilities because of decreased length of acute hospital stay and an early discharge before full functional recovery.¹ Therefore, the organization of PAC for this population becomes a major concern.²

A high percentage of patients are reported to suffer from delirium or delirium symptoms on admission to PAC.³, ⁴, ⁵ One study analyzed nursing home staff ratings of delirium symptoms that are listed on the Minimum Data Set and reported that 23% of newly admitted PAC patients had one or more delirium symptoms.³ It is important to note that these data were collected as part of a research project and may not reflect how staff detect delirium symptoms in clinical practice. Two other studies used a validated assessment implemented by research staff (ie, Confusion Assessment Method, CAM)⁶ to identify delirium in patients newly admitted to PAC and reported a delirium prevalence rate of 14% and 16% respectively.⁴, ⁷

Subjects admitted to a PAC facility experiencing delirium have poorer outcomes than subjects without delirium. Marcantonio et al⁵ reported that delirious PAC patients compared with nondelirious patients have a higher risk for rehospitalization (30% versus 13%), less chance to be discharged to the community (30% versus 73%), and higher 6-month mortality (25% versus 5.7%). Belleli and colleagues⁸ have also shown how delirium superimposed on dementia was predictive of a 2-fold increase in mortality in a population admitted to PAC rehabilitation setting. Kiely et al⁷ reported that patients whose delirium resolved by 2 weeks without recurrence, regained 100% of their prehospital function level, whereas patients who never resolved delirium retained less than 50% of their prefunctional level. This latter study underscores the potential importance of early detection of delirium, multicomponent evaluation, and treatment.

There are currently no data available describing how often PAC providers detect delirium or symptoms associated with delirium in either newly admitted patients or patients who continue to receive care in the PAC. However, it is probable that delirium and associated symptoms are frequently missed in PAC because there is evidence that delirium is significantly undetected in both hospital and emergency department settings.⁹, ¹⁰ One reason for the underdetection is that clinical staff may describe patient symptoms in the chart with terms that do not include the word delirium. In this regard, 2 studies used a validated chart-based method to detect delirium in a hospital setting.¹¹, ¹² One study¹¹ identified patients for whom clinical staff used keywords in the chart to describe behaviors that might be associated with delirium (eg, disorientation) and determined the relationship between these descriptions and delirium assessed with the CAM. Delirium was present in 15% of the patients according to the CAM and the chart-based key terms (any one descriptive term) identified delirium with a 74% sensitivity and 83% specificity. A second study¹² reported a 64% sensibility and an 85% specificity of the chart-based review in detecting delirium compared with the Confusion Assessment Method for the intensive care unit (CAM-ICU),¹³ with an overall agreement between chart method and the CAM-ICU of 72%.

The current study had 2 primary objectives: (1) to document the prevalence of keywords associated with delirium detected by clinical staff in PAC patients according to a chart review “keywords” list validated for delirium in the hospital setting and to detect the symptoms of delirium recorded on the Minimum Data Set; and (2) to document the actions taken by providers when patient behaviors were described with the delirium keywords list.

Back to Article Outline

Methods 

The design was a retrospective electronic and paper chart-based review study. Potential participants were patients admitted to 2 postacute facilities affiliated with Vanderbilt University Medical Center, Nashville, TN, from March 1, 2007, through March 31, 2008. An electronic search through the charts looking for 10 keywords associated with delirium, as previously validated in a chart-based review for delirium identification¹¹ (ie, confusion, disorientation, altered mental status, delirium, agitation, inappropriate behavior, mental status change, inattention, hallucination, lethargy) was performed. Severe dementia (a cognitive performance score of 5 to 6 as described below)¹⁴ was considered an exclusion criterion for the purpose of the study because it is more difficult for staff to discriminate behaviors that might be associated with delirium as opposed to severe dementia in this population. The Institutional Review Board at Vanderbilt University Medical Center, Nashville, TN, approved this study. Consent was waived because of the nature of the study.

A complete review of the medical records (electronic and paper charts) was performed by 2 geriatricians (A.M., L.S.). The medical record review took 30 to 45 minutes per patient.

Demographic data (age, sex, education, race), comorbidities (Charlson Comorbidity Index score),¹⁵ main discharge diagnosis from antecedent hospitalization organized by system, and length of stay were recorded. Hearing and vision impairment was recorded from the admission Minimum Data Set (MDS) version 2.0. Baseline cognitive status was obtained by the MDS Cognitive Performance Scale¹⁴ at admission and defined as no impairment (score 0), mild impairment (score 2), moderate impairment (score 3 to 4), and as severe impairment (score of 5 to 6). Diagnosis of depression was recorded from the charts. Functional status was assessed using the activities of daily living (ADLs) based on the ADLs summary scale obtained from the MDS (at discharge).¹⁶ Three stages were defined: early loss (extensive assistance in dressing, personal hygiene), middle loss (extensive assistance in toilet use, transfer, locomotion), and late loss (extensive assistance in bed mobility and eating).

The source of the information for the keywords (primary nurse, other nurse, primary physician, and other physician) was recorded and any actions following the identification of the keywords. Actions could include assessment, pharmacologic, or nonpharmacologic treatments.

Moreover, following a previous report of detection of delirium symptoms in a postacute setting,³ the MDS data on admission, at 14 days (when present), and at discharge were analyzed to evaluate the recognition of delirium symptoms. The MDS³ includes a section to evaluate 6 indicators of delirium: (1) easily distracted, (2) periods of altered perception or awareness of surroundings, (3) episodes of disorganized speech, (4) episodes of restlessness, (5) periods of lethargy, and (6) mental function varies over the course of the day. Each symptom is coded as 0 (behavior not present), 1 (behavior present, not of recent onset), 2 (behavior present, over last 7 days appears different from resident's usual functioning). For the purpose of the study, as previously indicated,³ patients were defined to have delirium symptoms if they had a score of 1 or 2 on any of the 6 delirium indicators on any MDS assessment.

Finally, we have collected mortality information during the PAC stay, and a 1-year follow-up for a subset of 25 patients discharged to the Nursing Homes (NHs) affiliated with the PAC facilities where this study was conducted.

Data characterizing the study population, the identification of keywords associated with delirium, and the actions following were examined by descriptive statistics. Independent sample T-test was used to examine the differences in the keywords between the group that received any actions and the one that had not. All analyses were performed using the SPSS version 15.0 (SPSS Inc, Chicago, IL).

Back to Article Outline

Results 

A total of 92 (10%) patients of 895 charts reviewed were identified whose charts contained keywords suggestive of delirium. Twelve of these patients were excluded for severe dementia and hence prevalence of keywords associated with delirium for patients not severely demented was estimated to be 9% (80 patients). The characteristics of the population are described in Table 1. Among 80 patients included in the study 44% had a high school education. Almost half of the studied population (44%) had a diagnosis of dementia; in particular, 20% of the patients had mild dementia and 24% had moderate dementia. The evaluation of the functional status at discharge showed an early functional loss in 48% of the patients, middle functional loss in 35%, and late functional loss in 13%. Of the 80 patients, 51 (64%) were discharged home, 25 (31%) were discharged to a nursing home setting, and 4 (5%) died in the PAC setting.

Table 1. Characteristics of the Study Population

Variables	(N = 80)
Age, mean ± SD	83 (±7)
Gender (female), n (%)	51 (64)
Dementia at baseline, n (%)∗	35 (44)
Depression at baseline, n (%)†	27 (34)
Functional Impairment, n (%)‡	76 (95)
Charlson Index, mean ± SD	7 (±2)
Hearing impairment, n (%)§	23 (28)
Vision impairment, n (%)	27 (33)
Main diagnoses reported by system, n (%)‖
Musculoskeletal	44 (55)
Cardiovascular	13 (16)
Respiratory	11 (14)
Neurologic	9 (11)
Gastrointestinal	3 (4)
Length of stay in postacute setting, mean ± SD	32.1 (±20.0)

Keywords suggestive of delirium were found in 9% of 883 charts of patients who were not severely demented. These keywords were noted by the primary nurse (99%), other nurse (8%), primary physician (69%), and other physician (1%). The different keywords recorded through the chart review are listed in Table 2, and it is clear that the term confusion (95%) was most frequently used to describe the resident's mental status. The specific term delirium was used in only 6 (7%) of the cases. The prevalence of delirium symptoms cumulatively from admission through discharge, as detected by the MDS, was 3% overall.

Table 2. Keywords Associated with Delirium Identified in the Chart Review¹¹

When the keywords associated with delirium were identified in the chart and patients referred to the physician (n = 55 patients or 69% of all patients), an action or intervention followed in all 55 patients as described in Table 3. Pharmacological interventions were more frequent (55, 100%), followed by assessments (38, 69%), transfer to emergency department (ED) and rehospitalized (19, 34%), and nonpharmacological interventions (11, 20%). In 12 (22%) patients, psychoactive medications (ie, opioids, benzodiazepines, selective serotonin reuptake inhibitors, atypical antipsychotics, sleep medications) were prescribed without performing diagnostic testing. The patients transferred to the ED and rehospitalized (24%) received an evaluation (ie, blood analysis, urine analysis, stool analysis) (73%) and pharmacological (100%) and nonpharmacological (15%) interventions in the postacute care setting before being rehospitalized. The most common reasons noted for hospital transfer were infection (68%), respiratory failure (16%), and acute anemia (16%). None of the patients who were not referred to the physician was transferred to the ED.

Table 3. Actions Following Recognition of Keywords Associated with Delirium in the Postacute Setting^∗

Actions Following Recognition of Keywords	(N = 55)
Pharmacological	55 (100)
Antibiotics	16 (29)
Benzodiazepines†	12 (22)
Atypical antipsychotics	8 (14)
Antidepressants‡	9 (16)
Opioids	7 (12)
Anticholinesterase	5 (9)
Sleep medicine	1 (2)
Nonpharmacological	11 (20)
Behavioral interventions§	9 (16)
Physical restraints	2 (4)
Assessments	38 (69)
Blood analysis	19 (34)
Urine analysis	11 (20)
Chest x-ray	6 (11)
Stool analysis	2 (4)
Transfer to ED‖and rehospitalized	19 (34)

The patients included in the “action taken” group who were referred to a physician had a higher number of keywords recorded (2.0 ± 0.9, P = .001) in the chart than the group who was not referred to a physician (1.0 ± 0.3). The word “confusion” was reported at a similar rate among the “action taken” group versus those not brought to the attention of the physicians (96% versus 92%). Moreover, the patients included in the “action taken” group generally (but not significantly) had a more “impaired” or vulnerable demographic and comorbidity profile including the facts that they were older (83 ± 7 versus 81 ± 7), with middle (40% versus 24%) and late functional loss (16% versus 4%), with hearing (35% versus 16%) and vision (36% versus 28%) impairment, and more depression (36% versus 28%) and dementia (51% versus 28%). In particular, the action-taken group presented higher prevalence of mild (22% versus 16%) and moderate dementia (29% versus 12%) compared with the group of patients for which the physician was not notified.

One of the patients not referred to the physician died during the PAC stay. Moreover, comparing patients in the “action-taken group” with the group not reported to the physician 31 (56%) versus 20 (80%) were discharged home, 21 (38%) versus 4 (16%) were discharged to a nursing home and 3 (5%) versus 1 (4%) died in the PAC.

It is worth noting that the total mortality for the subset of the 25 patients discharged to the nursing homes was 52%: 8% died between 1 and 3 months, 5% between 4 and 7 months, 4% from 8 months to 1 year.

Back to Article Outline

Discussion 

Health care providers (primarily nurses) described patient behaviors in the chart that are potentially associated with delirium in 9% of PAC patient charts who were not severely cognitively impaired but seldom used the term delirium. The term confusion was used in 95% of the cases to describe mental status when any word to describe mental status was used, whereas the term delirium was used in 7% of the cases when a descriptive word was provided in the chart. The patients in the “action-taken” group had more keywords charted than the other group, suggesting more severe symptoms of an acute change of a mental status. This has also been noted in a previous study¹⁷ in which nurses had a higher rate of documenting delirium symptoms when the patient displayed a more visible symptom profile. The 9% prevalence of the keywords associated with delirium is lower than the 14% to 16% delirium prevalence rate reported in PAC patients when standardized delirium assessments were used.⁴, ⁷

The chart-based review method¹¹ could potentially provide a method to screen for possible delirium, although delirium misclassification is probably an issue because this method was validated in the acute care setting. The prevalence of dementia is higher in the PAC setting than in acute care, and the word confusion could fail to distinguish delirium, dementia, or delirium superimposed on dementia. The underdetection of delirium is not surprising because this omission has been reported across health care settings.

Inouye et al¹¹ reported that the chart-based keyword method was most likely to miss delirium in patients who had cognitive impairment, higher sickness acuity, and high baseline risk of delirium. These characteristics are commonly found in the PAC population. Voyer et al¹⁷ reported, in a population of delirious patients actively evaluated with a validated assessment method, that at least one delirium symptom (eg, confusion, level of consciousness, perceptual disturbances, psychomotor agitation, disorientation, and memory impairment) was present in 64% of the nursing notes.

Patients not referred to the physicians compared with the “action-taken” group were more often discharged home (80% versus 56%). This might be because they had fewer symptoms associated with delirium and these symptoms may not have persisted at the discharge time.

Physicians reacted in all cases when notified about the presence of keywords in the chart with either orders for further assessments or treatments even though the treatments in some cases were not indicated if true delirium was present (eg, restraints, benzodiazepines) and in 12 patients (22%) psychoactive medications (ie, opioids, benzodiazepines, antidepressants, atypical antipsychotics, sleep medications) were prescribed without performing diagnostic tests. These latter data raise questions regarding the appropriateness of the pharmacological approach to patients with keywords potentially associated with delirium in this setting and should be more thoroughly evaluated in future studies. However, an important issue is also that physicians were not alerted to the keywords and no action was taken in 33% of the cases. The fact that physicians are reporting fewer keywords associated with delirium or delirium diagnosis than the nurses might be related to the frequency of clinical evaluation performed by physicians in the PAC setting where the study was conducted. In fact, physicians typically can be reimbursed to see a patient once per month or as needed if there is a change in condition, compared with the daily evaluation carried out by the nurses.

These latter data suggest 2 directions for future work

A recent review¹⁹ reported that nurses often do recognize that patients are confused and exhibiting inappropriate behavior but do not use these observations to move toward formal documentation of delirium. Currently in the PAC setting there is an average nurse-patient ratio of 1:20, hindering a full diagnostic delirium evaluation because of lack of time. The use of a method permitting the most efficient identification of delirium is thus clinically necessary.

In addition to the fact that this study is a retrospective pilot investigation, the study presents other limitations. First, the Chart Review Method as previously highlighted has not been validated in a PAC setting. Second, these data may not be generalizable to other PACs because this pilot project included only 2 sites. Third, it is possible that some reporting of symptoms to the physicians was done in a verbal manner and not documented, which would lead to an underestimate of how frequently physicians were notified about behaviors associated with delirium.

In conclusion, to the best of our knowledge, this is the first study to describe how PAC providers document behaviors associated with delirium. These data suggest that reviewing charts for keywords may identify high-risk patients for whom a more formal delirium assessment should be completed, and finally that delirium preventive or treatment interventions should be evaluated in the PAC setting.

Back to Article Outline

Acknowledgments 

The authors acknowledge with gratitude Nona Cooper, Brenda Ray, Melinda Newell, and Jamie Spicer, GNP, for the great support in accessing the electronic and paper charts.

Back to Article Outline

References 

1. Kauh B, Polak T, Hazelett S, et al. A pilot study: Post-acute geriatric rehabilitation versus usual care in skilled nursing facilities. J Am Med Dir Assoc. 2005;6:321–326
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (94 KB)
   • CrossRef
2. Young J, Green J, Forster A, et al. Postacute care for older people in community hospitals: a multicenter randomized, controlled trial. J Am Geriatr Soc. 2007;55:1995–2002
   • View In Article
   • CrossRef
3. Marcantonio ER, Simon SE, Bergmann MA, et al. Delirium symptoms in post-acute care: prevalent, persistent, and associated with poor functional recovery. J Am Geriatr Soc. 2003;51:4–9
   • View In Article
   • MEDLINE
   • CrossRef
4. Kiely DK, Bergmann MA, Murphy KM, et al. Delirium among newly admitted postacute facility patients: prevalence, symptoms, and severity. Gerontol A Biol Sci Med Sci. 2003;58:M441–M445
   • View In Article
5. Marcantonio ER, Kiely DK, Simon SE, et al. Outcomes of older people admitted to postacute facilities with delirium. J Am Geriatr Soc. 2005;53:963–969
   • View In Article
   • MEDLINE
   • CrossRef
6. Inouye SK, van Dyck CH, Alessi CA, et al. Clarifying confusion: the confusion assessment method. A new method for detection of delirium. Ann Intern Med 199015;113:941–948.
   • View In Article
7. Kiely DK, Jones RN, Bergmann MA, et al. Association between delirium resolution and functional recovery among newly admitted postacute facility patients. J Gerontol A Biol Sci Med Sci. 2006;61:204–208
   • View In Article
   • MEDLINE
   • CrossRef
8. Bellelli G, Frisoni GB, Turco R, et al. Delirium superimposed on dementia predicts 12-month survival in elderly patients discharged from a postacute rehabilitation facility. J Gerontol A Biol Sci Med Sci. 2007;62:1306–1309
   • View In Article
   • CrossRef
9. Kakuma R, du Fort GG, Arsenault L, et al. Delirium in older emergency department patients discharged home: Effect on survival. J Am Geriatr Soc. 2003;51:443–450
   • View In Article
   • MEDLINE
   • CrossRef
10. Hustey FM, Meldon SW. The prevalence and documentation of impaired mental status in elderly emergency department patients. Ann Emerg Med. 2002;39:248–253
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (74 KB)
    • CrossRef
11. Inouye SK, Leo-Summers L, Zhang Y, et al. A chart-based method for identification of delirium: Validation compared with interviewer ratings using the confusion assessment method. J Am Geriatr Soc. 2005;53:312–318
    • View In Article
    • MEDLINE
    • CrossRef
12. Pisani MA, Araujo KL, Van Ness PH, et al. A research algorithm to improve detection of delirium in the intensive care unit. Crit Care. 2006;10:R121
    • View In Article
    • CrossRef
13. Ely EW, Inouye SK, Bernard GR, et al. Delirium in mechanically ventilated patients: Validity and reliability of the confusion assessment method for the intensive care unit (CAM-ICU). JAMA. 2001;286:2703–2710
    • View In Article
    • MEDLINE
    • CrossRef
14. Morris JN, Fries BE, Mehr DR, et al. MDS Cognitive Performance Scale. J Gerontol. 1994;49:M174–M182
    • View In Article
    • MEDLINE
15. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–619
    • View In Article
    • MEDLINE
    • CrossRef
16. Morris JN, Fries BE, Morris SA. Scaling ADLs within the MDS. J Gerontol A Biol Sci Med Sci. 1999;54:M546–M553
    • View In Article
    • MEDLINE
    • CrossRef
17. Voyer P, Cole MG, McCusker J, et al. Accuracy of nurse documentation of delirium symptoms in medical charts. Int J Nurs Pract. 2008;14:165–177
    • View In Article
    • CrossRef
18. Flaherty JH, Rudolph J, Shay K, et al. Delirium is a serious and under-recognized problem: Why assessment of mental status should be the sixth vital sign. J Am Med Dir Assoc. 2007;8:273–275
    • View In Article
    • Full Text
    • Full-Text PDF (71 KB)
    • CrossRef
19. Steis MR, Fick DM. Are nurses recognizing delirium? A systematic review. J Gerontol Nurs. 2008;34:40–48
    • View In Article
    • CrossRef
20. Inouye SK, Bogardus ST, Charpentier PA, et al. A multicomponent intervention to prevent delirium in hospitalized older patients. N Engl J Med. 1999;340:669–676
    • View In Article
    • MEDLINE
    • CrossRef