HomeLarge Type Edition
HOME ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Services
Right arrow Download to citation manager
PubMed
Right arrow PubMed Citation
The Journals of Gerontology Series A: Biological Sciences and Medical Sciences 59:M344-M349 (2004)
© 2004 The Gerontological Society of America

Characteristics Associated With Delirium Persistence Among Newly Admitted Post-Acute Facility Patients

Dan K. Kiely1,, Margaret A. Bergmann1, Richard N. Jones1, Katharine M. Murphy1, E. John Orav4 and Edward R. Marcantonio1,2,3,5

1 Research and Training Institute
2 Department of Medicine, Hebrew Rehabilitation Center for Aged, Boston, Massachusetts.
3 Division of Gerontology, Beth Israel Deaconess Medical Center, Boston, Massachusetts.
4 Division of General Medicine, Brigham and Women's Hospital, Boston, Massachusetts.
5 Division on Aging, Harvard Medical School, Boston, Massachusetts.

Address correspondence to Dan K. Kiely, MPH, MA, Hebrew Rehabilitation Center for Aged, Research and Training Institute, 1200 Centre Street, Boston, MA 02131. E-mail: kiely{at}mail.hrca.harvard.edu


   Abstract
Top
 Abstract
Methods
 Results
 Discussion
 References
 
Background. Delirium is costly, common, and may persist for weeks or months. Therefore, the adverse impact of delirium on loss of independence may occur in the post-acute setting rather than in the hospital. The purpose of this study is to describe the rate of delirium persistence and identify baseline patient characteristics that are associated with delirium persistence at 1 month among newly admitted post-acute facility patients who were admitted with delirium.

Methods. Patients were recruited from 4 Boston area skilled nursing facilities specializing in post-acute care (PAC). Assessment instruments included the Confusion Assessment Method Diagnostic Algorithm, the modified Delirium Symptom Interview, the Memorial Delirium Assessment Scale (MDAS), and the Blessed Dementia Rating Scale (BDRS). Multiple logistic regression analyses were used to identify patient characteristics associated with delirium persistence (at 1 month).

Results. Nearly 51% of the 85 delirious patients enrolled in this study had delirium at their 1-month follow-up assessment. Four patient factors associated with delirium persistence were identified: older age (>=85 years), severe delirium at PAC admission (MDAS score >15), prehospital cognitive impairment based on proxy report [BDRS], and the presence of all 8 modified Delirium Symptom Interview symptoms at PAC admission. Our model has very good predictive power (area under the receiver operating characteristic = 0.85).

Conclusions. Delirium is persistent in the post-acute setting. If verified in further research, the risk factors found in this study could be used to identify patients who are likely to have delirium after 1 month, and may prove useful in developing and targeting interventions of care.

DELIRIUM is a common, morbid, and costly condition among hospitalized patients, which is associated with adverse events that lead to loss of independence (1–6). There is mounting evidence that delirium may persist for weeks or months (6–8). Furthermore, there is an accelerating trend to discharge patients quickly from acute care facilities. Many of these elders are discharged to post-acute care (PAC) facilities due to incomplete resolution of cognitive and functional problems that prevent their immediate return home. Consequently, much of the long-term sequelae of delirium may occur in the post-acute setting rather than the hospital.

Very little is known about delirium in the post-acute setting. We previously reported the prevalence and persistence of delirium symptoms among post-acute patients. (9) However, that study used secondary data, and assessments were performed by facility staff not specifically trained to detect delirium. More recently, we reported the prevalence, symptoms, and severity of delirium among newly admitted post-acute facility patients using assessments performed by interviewers specifically trained to detect delirium (10). This study did not follow these patients over time.

We know of no publications that focus on characteristics associated with delirium persistence in newly admitted post-acute facility patients. Given the morbid and costly consequences of delirium, the ability to identify patient characteristics associated with delirium persistence may be useful to post-acute clinical staff, discharge planners, and administrators.

Therefore, the purpose of this study is to describe the rate of delirium persistence and identify baseline patient characteristics that are associated with delirium persistence at 1 month among newly admitted post-acute facility patients who were admitted with delirium.


   METHODS
Top
 Abstract
 Methods
Results
 Discussion
 References
 
Study Population
Patients were recruited between October 1, 2000 and February 14, 2002 as part of a randomized trial of a Delirium Abatement Program from 7 active Boston area skilled nursing facilities (SNF) specializing in PAC. Only patients receiving care in 4 facilities randomly assigned to the control group were included in this study because we did not want the potential effects of the intervention to influence the delirium rate measured at the follow-up assessment. All patients in this study had delirium at post-acute admission (baseline), and their delirium status at the 1-month assessment was the dependent variable in analyses. In addition to having delirium at admission, eligible patients in this study were aged 65 or older and admitted directly from an acute-care facility. They also spoke English, did not have a serious hearing impairment, were communicative prior to acute illness, were not admitted for terminal care (life expectancy < 6 months), and lived within a 25-mile radius of our research site. Interviews were conducted by trained research assistants and completed preferably within 72 hours (average time to interview = 2.5 days), but not longer than 5 days after admission. The 1-month interview was defined as 28 (± 3) days beyond the date of admission. Multiple assessors were employed, but the same assessor usually performed both baseline and follow-up interviews. The interrater reliability was excellent (kappa = 0.90) (11).

Dependent Variable
The Confusion Assessment Method (CAM) is a diagnostic algorithm derived from DSM-III-R criteria for delirium that has been validated against a psychiatrist's diagnosis. The CAM allows trained research assistants to diagnose delirium with greater than 95% sensitivity and specificity, even in populations with a high prevalence of dementia (12). The CAM diagnostic algorithm involves 4 criteria: 1) an acute change in mental status with a fluctuating course, 2) inattention, 3) disorganized thinking, and 4) an altered level of consciousness (12). Using the CAM algorithm, delirium was considered present if CAM criteria 1 and 2 were present, and either of criteria 3 or 4 was present.

Independent Variables
The Delirium Symptom Interview (DSI) (13) is a valid and reliable structured interview for diagnosing the presence of specific critical symptoms of delirium in an objective and straightforward manner, and can be administered by lay interviewers. Consistent with the current definition of delirium, modifications were made in the original DSI instrument to separate the symptom of "disturbance of attention" from that of "disturbance of consciousness." Items addressing informal observations of inattention were redirected as triggers to the presence of inattention, a new eighth symptom. Additionally, an item was added to incorporate and rate performance on formal tasks of attention. Formal tasks of attention included days of the week and months of the year backwards along with digit span testing up to five digits forward and four digits backwards. This instrument was administered at admission to the post-acute facility.

A variable representing whether the patient had all 8 modified DSI items (referred to as mDSI All 8 Symptoms) was created and used in analyses. The following 8 items were included: 1) disorientation, 2) disturbance of sleep, 3) perceptual disturbance, 4) disturbance of attention, 5) disturbance of consciousness, 6) incoherent speech, 7) abnormal psychomotor activity, and 8) fluctuating behavior.

The Memorial Delirium Assessment Scale (MDAS) (14) allows trained research personnel to quantify the severity of delirium based on 10 features, each scored from 0 to 3 for a maximum score of 30. The 10 MDAS features include reduced level of consciousness, disorientation, short-term memory impairment, impaired digit span, reduced ability to maintain and shift attention, disorganized thinking, perceptual disturbance, delusions, decreased or increased psychomotor activity, and sleep–wake cycle disturbance. A higher score indicates greater delirium severity. The MDAS provides an explicit description of the criteria for scoring, can be completed in 5 minutes, and integrates behavioral observations with objective cognitive testing. It was completed at admission in the post-acute facility. A two-category delirium severity variable was created as either mild-moderate (<=15 points) or severe (>15 points) based on the distribution of MDAS scores in this population.

The Blessed Dementia Rating Scale (BDRS) (15) was used to assess cognitive impairment prior to hospitalization. This scale was designed to be completed by the patient's primary caretaker (proxy) and has been corroborated with pathological findings. It is measured at baseline and asks the proxy to assess the patient's cognitive ability prior to hospitalization. Scores range from 0–28, with higher scores indicating greater levels of impairment. We defined cognitive impairment (using Blessed's recommendation of categorizing individual's performance on this measure) as either impaired (a score of >=4) or normal (a score of <4).

Gender and race were also included. Race was grouped as white and nonwhite because the vast majority of patients were white. Education level was grouped as either less than a high school education or at least a high school education. Age was categorized at >=85 years or <85 years.

Data on the patient's preillness functional status was obtained from the caregiver using the modified Activities of Daily Living (ADL) scale (16). The ADL scale remains a well-established and extremely useful scale for measuring recovery from acute illness in frail elders (17). The scale has been modified to include walking and grooming in addition to the 6 original activities described by Katz (bathing, dressing, toileting, continence, transferring, and feeding) (16). The modified ADL scale was used to measure premorbid ADL function prior to the acute illness, which resulted in hospitalization. Because of the compromised cognitive status of the participants, the ADL scale was administered to next of kin/proxy. Studies have generally found that functional data derived from proxies is comparable to self-report or performance-based measures (18). The values of this scale range from 0 (complete dependence) to 16 (independence). We created a functional dependence variable that was designated as dependent if a patient had an ADL value below the median score. This was done mainly because the scale was not normally distributed, but also because it is easier for the reader to comprehend the meaning of the effect measure (relative risk, RR) and its confidence interval (CI).

Comorbidity diagnoses have traditionally been underreported, so we utilized two different sources. One source of medical comorbidity was assessed using the Charlson Comorbidity Scale (CCS) (19), a validated scale commonly used in epidemiological research. A brief interview has been validated to obtain the data necessary to complete the Charlson index from patients or caregivers (20). This interview was administered to the proxy at study intake to assess preillness comorbidity. Because the distribution of this scale was not normally distributed, we categorized a variable (CCS high group) as greater than the median or equal/less than the median, and used this variable in analyses. The CCS item Alzheimer's disease and other dementia were also individually included as potential characteristics associated with delirium persistence.

We felt that patients with a primary or secondary diagnosis involving the central nervous system (CNS) might be less likely to resolve their delirium. Therefore, in addition to the CCS interview described above, we assessed for the presence of a CNS diagnosis using post-acute facility medical record reviews including hospital discharge summary, if available. A variable was created that represented whether the primary or secondary diagnosis for admitting the patient to the hospital was for a CNS problem (i.e., stroke, Parkinson's disease). Finally, we created a variable representing whether a CNS problem was present based on either the CCS or the primary (or secondary) diagnosis from the medical record.

Statistical Analyses
Chi-square and bivariable (one dependent variable and one independent variable) logistic regression analyses were initially performed to help identify potential risk factors for delirium persistence. Chi-square analyses were used to estimate the RRs and 95% CIs. (21) Variables that were significant in the bivariable models were entered into a stepwise logistic regression analysis. Variables that were significant in the stepwise model were included in the final multivariable logistic regression model. Odds ratios (ORs) and corresponding 95% CIs were calculated to estimate the association between specific patient characteristics and delirium persistence. The area under the receiver operating characteristic (ROC) curve (22,23) for the final multivariable logistic regression model was determined and used as an estimate of the strength of the model. Because the OR overestimates the true RR when the incidence of the outcome is not rare (>10%), we employed a commonly used adjustment method that more accurately estimates the true RR (24). An alpha level of 0.05 was used to determine statistical significance, and all analyses were performed using SAS software (SAS, Institute, Inc., Cary, NC) (25,26).


   RESULTS
Top
 Abstract
 Methods
 Results
Discussion
 References
 
Among the 2051 patient admissions in facilities randomized to the control group at the time of this analysis, 1676 (82%) were eligible. A total of 1240 (74%) were screened and 188 (15%) were delirious. Of the 188, 47 proxies refused, 28 proxies did not respond within the enrollment period (despite repeated efforts), and 2 died before the proxy was reached. A total of 111 patients (baseline) were enrolled in the study. Of the 111, 85 patients had a 1-month assessment of delirium. Of the 85, a total of 56 patients were interviewed in a PAC, 1 in a hospital, and 28 in a residential setting. Among the 26 who had a baseline assessment but not a 1-month follow-up assessment, 8 died during this period, 6 patients withdrew from the study, and the remaining 12 patients were temporarily unavailable at the 1-month assessment due to rehospitalization or illness.

The average age of patients was 85 years (standard deviation [SD] = 7 years), the median ADL score was 15, and the median CCS score was 3. Table 1 provides descriptive information on characteristics of the 85 patients with baseline and 1-month follow-up assessments who were included in the initial bivariable analyses. Delirium was present in more than half (51%) of the 85 patients completing a 1-month follow-up assessment.


View this table:
[in this window]
[in a new window]
  		Table 1. Descriptive Information on Delirium at 1 Month, and on Factors Potentially Associated With Delirium Persistence for Patients Assessed at Admission to a Post-Acute Care Facility (N = 85).

 
Table 2 presents RRs, 95% CIs, and p values associated with chi-square analyses representing patient characteristics considered potentially associated with delirium persistence. Eight of these characteristics were statistically significant in bivariable models: older age (>=85 years), high school education or more, premorbid functional dependence, cognitive impairment (prehospitalization), mDSI All 8 Symptoms (from admission mDSI), severe delirium (from admission MDAS), CNS condition (Charlson comorbidity diagnosis), and Alzheimer's disease (Charlson comorbidity diagnosis).


View this table:
[in this window]
[in a new window]
  		Table 2. Bivariable Relative Risks and 95% Confidence Intervals for Specific Patient Characteristics (N = 85).

 
Table 3 presents the results of the final multivariable logistic regression analysis. Four patient characteristics were included in this model: mDSI All 8 Symptoms, cognitive impairment (prehospitalization), severe delirium, and older age (>=85 years). The model performance was excellent (area under the ROC = 0.85). Because the incidence of the outcome (delirium persistence) is high (51%), the ORs overestimate the true RRs. Consequently, the Zhang and Yu adjusted RR and 95% CI estimates were also included in Table 3. These estimates provide a more accurate estimate of the true RRs. Each of these risk factors for delirium persistence imparts an increased risk of delirium persistence ranging from 60% to 80% (Table 3).


View this table:
[in this window]
[in a new window]
  		Table 3. Odds Ratios and 95% Confidence Intervals, and Relative Risks and 95% Confidence Intervals for Specific Patient Characteristics Included in the Final Multivariable Logistic Regression Analysis (Area Under the Receiver Operating Characteristic Curve =.85).

 
None of the 10 patients (12%) free of all 4 risk factors at PAC admission had persistent delirium at 1 month. Among the 33 patients (39%) who had 1 risk factor, more than 27% showed signs of persistent delirium. Of the 42 patients (49%) who had 2 or more risk factors, 81% showed signs of persistent delirium.


   DISCUSSION
Top
 Abstract
 Methods
 Results
 Discussion
References
 
In this study of 85 enrolled post-acute patients who were admitted directly from acute-care hospitals and were delirious at post-acute admission, 4 patient characteristics associated with delirium persistence at the 1-month assessment were identified in a multivariable model: 1) the presence of all 8 delirium symptoms (mDSI) (post-acute admission), 2) cognitive impairment (prehospitalization), 3) severe delirium (post-acute admission), and 4) older age (>=85 years). The predictive model containing these 4 characteristics performed very well.

More than half of the patients in this study had delirium at the 1-month follow-up assessment. This finding supports mounting evidence that a substantial percent of delirium cases persist for 1 month or more. The persistence of delirium has been studied in other populations. Levkoff and colleagues (7) reported that 96% of the 125 hospitalized patients in their study did not experience complete resolution of their delirium before discharge from the hospital. Furthermore, after hospital discharge, approximately 79% of patients at 3 months and 82% of patients at 6 months did not resolve all new delirium symptoms experienced in the hospital. Marcantonio and colleagues (6) studied delirium in a population of 126 elderly people admitted emergently for surgical repair of hip fracture and reported that delirium persisted for at least 1 month in one third of the patients. Kelly and colleagues (27) studied 214 hospital admissions and reported that 72% of patients who survived the hospitalization showed persistent delirium at the time of discharge. Delirium was persistent in the 55% of surviving patients at 1 month, and 25% of surviving patients at 3 months. Marcantonio and colleagues (9) studied 551 patients aged 65 years and older newly admitted to post-acute facilities (SNFs and rehabilitation hospitals) from acute care hospitals. They reported that only 14% showed complete resolution of delirium symptoms after 1 week. They also reported that 64% of patients with delirium symptoms at admission had the same number or more delirium symptoms 1 week later.

We know of no publications that have focused on patient characteristics associated with delirium persistence in a post-acute population, or in any population. However, some of the patient characteristics associated with delirium persistence in this publication have been reported to be associated with incident delirium in a hospitalized population. Levkoff and colleagues (7) reported that older age (>80 years) and preexisting cognitive impairment were associated with delirium incidence in both community and institutionalized populations. Marcantonio and colleagues (28) studied 1341 patients 50 years or older admitted for major elective noncardiac surgery and reported, among other characteristics, that older age (>=70 years) and poor cognitive status were independent correlates of postoperative delirium. Inouye and colleagues (29) studied 107 hospitalized patients aged 70 years or older who did not have dementia or delirium at admission and reported, among other characteristics, that cognitive impairment was a risk factor for delirium.

This study has advantages and limitations worthy of discussion. Though this study used an established and validated diagnostic algorithm (CAM) and trained research personnel to assess delirium at baseline and 1-month follow-up, we cannot be sure of the delirium status of every patient throughout the 1-month follow-up. Possibly some of these patients could have resolved their delirium during this time, say at 1 week, and then later redeveloped their delirium as identified by the 1-month assessment. Our data was collected from a single metropolitan region, but included 4 facilities. The prevalence of delirium may vary in cities with less pressure to discharge patients from hospitals, or at PAC facilities that admit fewer acute patients. Furthermore, the results of our study involving patients in a PAC skilled nursing facility may not generalize to individuals receiving PAC in a residence or rehabilitation hospital. We only considered factors assessed at the time of PAC admission and acknowledge that there may be factors assessed after PAC admission that may be associated with delirium persistence. Also, there are several delirium incidence risk factors (i.e., vision, hearing, dehydration) that we did not include in our analyses because we did not obtain information on these factors or these factors were part of our exclusionary criteria. Finally, although our assessments were performed within 2.5 days of admission, we cannot be sure if some of the patients developed delirium after they were admitted to the PAC facility.

Concerns of collinearity between the delirium severity and mDSI All 8 Symptoms variables motivated us to examine this relationship. A cross-tabulation of these two variables revealed that only 4 of the 85 patients were positive on both variables and 59 patients were negative on both variables. Eight patients did not have severe delirium but did have all 8 DSI symptoms. Fourteen patients had severe delirium but did not have all 8 DSI symptoms. These results gave us assurance that these were separate constructs. We believe that the delirium severity variable measures the "depth" of delirium (the total severity of delirium symptoms experienced by the patients). In contrast, the mDSI all 8 symptoms variable measures the "breadth" of delirium (whether the patient experienced all the major symptoms of delirium, regardless of their severity). As such, these represent both theoretically and empirically different constructs.

Conclusion
The majority of post-acute patients did not resolve their delirium after 1 month. Four patient characteristics were identified that can be used to estimate who will have delirium after 1 month. This study adds to a growing body of evidence suggesting that delirium prevalence is relatively high in the post-acute setting and cases of delirium are often persistent over time. Further research is needed in post-acute populations to determine if patient characteristics identified in this study can be validated and used to predict patients with an elevated risk of having delirium after 1 month. These results may prove useful in developing and targeting interventions of care.


   Acknowledgments
 
Supported in part by a grant from the National Institute on Aging (RO1 AG 17649). Dr. Marcantonio is a recipient of a Paul Beeson Physician Faculty Scholarship in Aging Research.

The authors thank Monique Bussell, Kerry Clark, Kathryn Johnson, Maria Kereshi, Jennifer Kettell, Melissa McKenna, Mary Michaels, and Sara Van Valkenburg for their efforts to enroll and interview patients for this study, and registered nurses Judith Coulombre and Maryann Wallace for their efforts in completing medical record reviews.

The authors also acknowledge Ellen Gornstein and Pamela A. Heidell for reviewing this manuscript, and Aleksandra Brenckle for assistance preparing this manuscript for submission.

Received September 30, 2003

Accepted December 30, 2003


   References
Top
 Abstract
 Methods
 Results
 Discussion
 References
 

  1. Inouye SK, Foreman MD, Mion LC, Katz KH, Cooney LM, Jr. Nurses' recognition of delirium and its symptoms: comparison of nurse and researcher ratings. Arch Intern Med.. 2001;161:2467-2473.[Abstract/Free Full Text]
  2. Murray AM, Levkoff SE, Wetle TT, et al. Acute delirium and functional decline in the hospitalized elderly patient. J Gerontol A Biol Sci Med Sci.. 1993;48:M181-M186.
  3. Inouye SK. Delirium in hospitalized older patients: recognition and risk factors. J Geriatr Psychiatry Neurol.. 1998;11:118-125 discussion 157–158.
  4. O'Keeffe S, Lavan J. The prognostic significance of delirium in older hospital patients. J Am Geriatr Soc.. 1997;45:174-178.[Medline]
  5. Francis J, Kapoor WN. Prognosis after hospital discharge of older medical patients with delirium. J Am Geriatr Soc.. 1992;40:601-606.[Medline]
  6. Marcantonio ER, Flacker JM, Michaels M, Resnick NM. Delirium is independently associated with poor functional recovery after hip fracture. J Am Geriatr Soc.. 2000;48:618-624.[Medline]
  7. Levkoff SE, Evans DA, Liptzin B, et al. Delirium. The occurrence and persistence of symptoms among elderly hospitalized patients. Arch Intern Med.. 1992;152:334-340.[Abstract/Free Full Text]
  8. Manos PJ, Wu R. The duration of delirium in medical and postoperative patients referred for psychiatric consultation. Ann Clin Psychiatry.. 1997;9:219-226.[Medline]
  9. Marcantonio ER, Simon SE, Bergmann MA, Jones RN, Murphy KM, Morris J. Delirium symptoms in post-acute care: prevalent, persistent, and associated with poor functional recovery. J Am Geriatr Soc.. 2003;51:4-9.[Medline]
  10. Kiely DK, Bergmann MA, Murphy KM, Jones RN, Orav EJ, Marcantonio ER. Delirium among newly admitted postacute facility patients: prevalence, symptoms, and severity. J Gerontol Med Sci.. 2003;58A:M441-M445.
  11. Simon SE, Bergmann MA, Marcantonio ER. Reliability of a comprehensive delirium assessment utilizing four instruments. Gerontologist.. 2001;41:(Suppl): M365.
  12. Inouye SK, van Dyck CH, Alessi CA, Balkin S, Siegal AP, Horwitz RI. Clarifying confusion: the confusion assessment method. A new method for detection of delirium. Ann Intern Med.. 1990;113:941-948.
  13. Albert MS, Levkoff SE, Reilly C, et al. The delirium symptom interview: an interview for the detection of delirium symptoms in hospitalized patients. J Geriatr Psychiatry Neurol.. 1992;5:14-21.
  14. Breitbart W, Rosenfeld B, Roth A, Smith MJ, Cohen K, Passik S. The Memorial Delirium Assessment Scale. J Pain Symptom Manage.. 1997;13:128-137.[Medline]
  15. Blessed G, Tomlinson BE, Roth M. The association between quantitative measures of dementia and of senile change in the cerebral grey matter of elderly subjects. Br J Psychiatry.. 1968;114:797-811.[Abstract/Free Full Text]
  16. Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW. Studies of illness in the aged. The Index of ADL: a standardized measure of biological and psychosocial function. JAMA.. 1963;185:914-919.
  17. Landefeld CS, Palmer RM, Kresevic DM, Fortinsky RH, Kowal J. A randomized trial of care in a hospital medical unit especially designed to improve the functional outcomes of acutely ill older patients. N Engl J Med.. 1995;332:1338-1344.[Abstract/Free Full Text]
  18. Magaziner J, Zimmerman SI, Gruber-Baldini AL, Hebel JR, Fox KM. Proxy reporting in five areas of functional status. Comparison with self-reports and observations of performance. Am J Epidemiol.. 1997;146:418-428.[Abstract/Free Full Text]
  19. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis.. 1987;40:373-383.[Medline]
  20. Katz JN, Chang LC, Sangha O, Fossel AH, Bates DW. Can comorbidity be measured by questionnaire rather than medical record review? Med Care.. 1996;34:73-84.[Medline]
  21. Kahn H, Sempos C. Statistical Methods in Epidemiology. Oxford: Oxford University Press; 1989.
  22. Metz CE. Basic principles of ROC analysis. Semin Nucl Med.. 1978;8:283-298.[Medline]
  23. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology.. 1982;143:29-36.[Abstract/Free Full Text]
  24. Zhang J, Yu KF. What's the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA.. 1998;280:1690-1691.[Abstract/Free Full Text]
  25. SAS Institute. SAS/STAT User's Guide, Version 8. Cary, NC: SAS Institute, Inc.; 1999.
  26. SAS Institute. SAS Language Reference: Dictionary, Version 8. Cary, NC: SAS Institute, Inc.; 1999.
  27. Kelly KG, Zisselman M, Cutillo-Schmitter T, Reichard R, Payne D, Denman SJ. Severity and course of delirium in medically hospitalized nursing facility residents. Am J Geriatr Psychiatry.. 2001;9:72-77.[Medline]
  28. Marcantonio ER, Mangione CM, Goldman L, Orav EJ, Lee TH. Postoperative delirium is associated with poor outcome after non-cardiac surgery. J Am Geriatr Soc. 1993;41(SA1).
  29. Inouye SK, Viscoli CM, Horwitz RI, Hurst LD, Tinetti ME. A predictive model for delirium in hospitalized elderly medical patients based on admission characteristics. Ann Intern Med.. 1993;119:474-481.[Abstract/Free Full Text]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Services
Right arrow Download to citation manager
PubMed
Right arrow PubMed Citation

HOME ARCHIVE SEARCH TABLE OF CONTENTS