The Trouble With Delirium In AIDS Patients:

The study was supported by the Special Projects of National Significance (SPNS) Program, Health Resources and Services Administration, Grant No. BRU900126. The contents of this paper are soley the responsibility of The Measurement Group, and do not necessarily represent the views of the SPNS program.

The authors wish to thank Angela Powell, M.P.H., for her assistance on the project.

The following article is taken from AIDS Patient Care, Volume 10, Number 5, 1996, Pages 282-287. The article is reproduced courtesy of Mary Ann Liebert, Inc. Publishers, 2 Madison Avenue, Larchmont, NY 10538, 914.834.3100, Fax 914.834.3771.

Delirium is among the most common neuropsychiatric disorders seen in medically ill patients.¹ Its occurrence has been documented to range from 7%² to 57%³ in medically ill populations, but as high as 85% in terminally ill cancer patients.⁴ Despite the frequency of its occurrence, however, delirium is greatly under-recognized by primary care providers. One study indicated that primary care physicians recognized delirium in only 8 of 48 patients meeting clinical diagnostic criteria for the disorder.⁵ Another indicated that nurses reported symptoms of delirium in only 10% of patients with high acuity illness (e.g., terminally ill cancer patients, patients in critical care) even though delirium has been documented to occur in much higher percentages of this population.⁴

Little is known about the occurrence of delirium in patients with HIV infection. Only two research studies conducted in the United States address delirium in patients with HIV/AIDS.^6,7 In these prospective reports, between 30-57% of medically hospitalized patients with HIV/AIDS were diagnosed by psychiatrists as delirious. Non-psychiatric primary care providers, however, have been shown to greatly under-report delirium in AIDS-related medical admissions. Among 2,846 AIDS-related medical admissions to 49 hospitals in Washington State between 1990 and 1992, retrospectively collected chart review data showed only 2% (n = 49) of these admissions as having an episode of delirium during the course of hospitalization.⁸ Compared to the prospective psychiatric studies indicating a delirious episode in 30-57% of AIDS-related medical admissions, it is clear that delirium is under-reported in hospitalized HIV/AIDS patients by primary care providers.

These findings are troublesome, given that delirium has been associated with adverse patient outcomes such as decreased independence requiring institutionalization, persistent decreased cognitive functioning, and increased length of hospitalization.^9-12 Delirium has also been associated with death. One study indicated that 25% of medically ill delirious patients died during the index hospitalization.¹³ Another reported that 37% of patients with delirium died within one to three months of the onset of this serious disorder.¹⁴ Yet, delirium is a treatable condition. Clearly, the correct and early identification of delirium is of utmost importance.

The authors investigated the occurrence of delirium in a Washington State community-based skilled nursing facility (SNF). Approximately 120 to 150 persons live at this 35-bed SNF each year. The typical patients are white, single, gay men over the age of 30, admitted because of increased difficulties managing at home. The purpose of this paper is to document the occurrence of delirium, as measured subjectively and objectively, in this Washington State community-based SNF specializing in the care of terminally ill AIDS patients. The paper highlights discrepancies between the occurrence of delirium and health care providers' identification and documentation of this condition, as well as inconsistencies between the percentage of patients who might benefit from a psychiatric consultation, staff’s request for such a consult, and the number of psychiatric consults given. Barriers to delirium identification are presented, as well as the training needs of health care providers.

Several methods were employed to investigate the occurrence of delirium and psychiatric consultations, and providers' identification of delirious episodes in patients. Health care providers' perceptions of the occurrence of delirium were measured via a self-report, paper and pencil, needs assessment survey. A retrospective review of patients' charts was conducted at the SNF to document the number of charts which included a diagnosis of delirium, the actual occurrence of delirium in these same patients as determined via a diagnostic algorithm, and the number of psychiatric consults ordered.

Content for the anonymous and voluntary learning needs assessment survey was a modified version of the Health Resources and Services Administration (HRSA) National AIDS Education Training Projects' national learning needs assessment survey implemented in 1994 - 1995. The list of learning issues was developed by an independent consulting firm in Atlanta, Georgia, in consultation with an expert working group. The working group included representatives from eight of the 17 funded AIDS Education and Training Centers across the country, and 22 staff from government agencies with an involvement or interest in services for people with HIV/AIDS and training of professionals to serve those with HIV/AIDS. The modified survey and its cover statement were reviewed by professional representatives at the SNF to assure face and content validity for the site. Both were revised per the facility's feedback prior to distribution. The survey and cover statement also passed University of Washington human subjects review before distribution.

The survey and cover letter were handed out to the SNF's 43 on-site professional staff-- predominantly nurses/aides, but including social workers, clergy, and occupational therapists in June, 1995. Multiple opportunities to complete the survey were given in order to assure a high percentage of full-time on-site staff receiving the survey. Completed surveys were either handed back to the study staff or to another designated person at the SNF. Physicians who either worked at or admitted/referred patients to the SNF during the year before the learning needs assessment was conducted received the survey by mail. However, as these physicians tended to work at or admit patients to a number of HIV/AIDS facilities in the area, their needs assessment answers were not specific to a particular site. Their results were not included in the needs assessment analyses for this report.

The survey items included background data which might influence training needs: e.g., primary functional role; year of completion of formal professional education; and number of hours of training regarding the neuropsychiatric aspects of HIV/AIDS in the past three years. Providers also rated their comfort level in providing care to different patient groups (i.e., patients with AIDS, delirium, and other neuropsychiatric conditions associated with an AIDS diagnosis), and advocating for a psychiatric consult for patients. The survey also asked providers the percentage of patients during the 6 months prior to completing the survey for whom they advocated a psychiatric consult, the barriers to doing this, the percentage of their current HIV/AIDS patients diagnosed with HIV/AIDS related neuropsychiatric conditions, and the percentage of patients diagnosed as delirious during their stay at the SNF. Finally, the survey asked providers to indicate the extent to which they felt they required training on a list of 25 neuropsychiatric diagnosis and treatment issues, of which 8 were specific to delirium: 1 = no/minor need; 2 = moderate need; 3 = major need. Learning need scores were calculated for each listed topic by assigning a score of 1 for no/minor need or missing value, 2 for moderate need, and 3 for major need: the score range possible for each topic was 33 to 99.

A retrospective chart review was conducted on the charts of 137 AIDS patients who were either discharged from or died at the SNF during 1994. A chart review form was developed to abstract demographic information, length of stay, residence at admission, discharge placement (if applicable), payment method, past and current medical (including psychiatric) conditions, and psychiatric consultation utilization during stay. The presence or absence of delirium symptoms was assessed using the Confusion Assessment Method (CAM)15, a diagnostic algorithm which helps identify the presence or absence of delirium. Interrater reliability checks were randomly performed on approximately 25% of the overall sample. To identify possible etiologic factors for the delirium, chart progress notes for the two days preceding the onset of delirium symptoms were reviewed.

Chi-square analyses were used to assess categorical variables. Differences between provider groups, nurses versus other providers, administrators versus direct care providers, were assessed using two-tailed t tests. Significance level was set at p < 0.05 for all analyses.

Thirty-three (77%) of the learning needs assessment surveys were completed and returned. The majority of returned surveys (73%) were from nurses (including nurse practitioners, registered nurses, nurse aides). The remainder were from allied and other health professional fields (e.g., occupational therapists, social workers). Eighty-eight percent of the survey respondents were direct care providers as opposed to administrators. The majority (63%) had completed their professional training more than 10 years ago. All but one staff person reported receiving some HIV/AIDS training in the past three years; 45% reported more than 50 hours of such training. For the same time period, approximately one-quarter (24%) reported receiving zero hours of neuropsychiatric training, but a significant percentage (42%) reported having between 1 and 10 hours.

Comfort levels in terms of providing care for persons with HIV and AIDS were high, with 97% reporting being either 'comfortable' or 'very comfortable' with both groups of patients. Comfort levels providing care for persons with HIV/AIDS-related neuropsychiatric conditions and delirium were lower: 63% and 66%, respectively, reported being either 'comfortable' or 'very comfortable'. Staff who returned the survey were most uncomfortable providing care for persons with a history of violence; 18% reported being either 'comfortable' or 'very comfortable'.

While the majority (76%) of the respondents reported being either 'comfortable' or 'very comfortable' ordering and/or advocating for a psychiatric consult, approximately half (55% ) actually did so for only a small percentage (0 - 10%) of patients in the 6 months prior to completing the survey. This percentage was not proportionate to the percentage of patients identified as being diagnosed with HIV/AIDS-related neuropsychiatric conditions: 54% estimated that more than half of their patients were neuropsychiatrically ill. An additional quarter (24%) did not know the percentage of their patients diagnosed with HIV/AIDS-related neuropsychiatric conditions. Barriers to ordering and/or advocating for a psychiatric consultation included: doctor opinion of need; the availability of psychiatric service; patient unwillingness; cost/insurance issues; not perceiving advocating for a psychiatric consult as part of their role; and lack of knowledge regarding delirium and/or how to determine the need for a psychiatric consult.

Provider estimates of the percentage of current AIDS patients diagnosed as delirious during their stay at the SNF ranged considerably--from 0% to 90%-- with an average of 36%, as estimated by the 29 direct care providers (excluding administrators and support staff). Specifically, 14% percent of the 29 direct care providers who returned the survey estimated that none of their current AIDS patients had been diagnosed as delirious during their stay. Nearly one-quarter (21%) estimated that 10% or less had been so diagnosed whereas 31% estimated that between one-quarter and one-half of their patients had been diagnosed with delirium.

With the scoring system applied to the training needs, the range of scores possible was 33 to 99; the higher the score, the greater perceived need for training. Actual training need scores achieved were 37 to 89. The top rated training need was stress counseling for staff (score of 89), followed by educating patients/families about delirium (86), managing symptoms associated with delirium (85), recognizing symptoms associated with delirium (81), and anxiety in persons with HIV/AIDS and delirium (80). All eight of the delirium-specific training issues listed on the survey were very highly rated by the providers in terms of need for training (scores between 78 and 86).

There was no significant difference between nurses and other providers or between direct care providers and administrative staff on hours of HIV/AIDS training, hours of neuropsychiatric training, and comfort levels with various patient groups. Neither were there significant differences between these groups in terms of their perceptions regarding the number of patients needing a psychiatric consult, estimated number of patients with neuropsychiatric illness, or estimated number of patients experiencing delirium at some point during admission.

Patient demographics in the charts reviewed were consistent with the SNF's patient population statistics over the past three years. Briefly, the median age of patients was 37 years (mean = 39 ± 9 years), with ages ranging from 25 to 76 years. The majority of patients were Caucasian, single/never married, and were exposed to HIV through male/male sex. Patients were admitted from a variety of sources, with the greatest majority arriving either from their homes (46%) or from hospitals (34%). Public insurance/assistance was the most common method of payment for two-thirds of the patients; the other third used private insurance. Once admitted, the median length of stay at the SNF was 54 days (mean = 78 ± 76 days), with stays of less than 1 day to 415 days. Approximately 88% of patients died at the facility, but 12% were discharged. Approximately half (56%) of these 16 discharged patients went to their homes; the remainder were discharged to medical units in hospitals (22%), and to unknown placements (22%).

As indicated in Table 1, the most frequently charted AIDS-associated conditions were HIV-encephalopathy (55%), wasting syndrome (55%), Mycobacterium avium complex (38%), non-central nervous system cytomegalovirus (35%), and non-central nervous system neurological problems (e.g., peripheral neuropathy, 34%). The retrospective chart review found that 74% of the patients were documented as having some psychiatric morbidity during their stay at the SNF; the most commonly documented conditions were dementia (55%) and mood disorders (22%).

The CAM retrospectively diagnosed 46% of the patients (n = 63) as having an episode of delirium at some point during their stay. Interrater reliability was good (k =0.78). Of these 63 patients, approximately 17% (n = 11) experienced two or more episodes. Delirium was evident in the first week of admission for 13% of all patients and even more common in patients (30%) in the days prior to death. Important to note is that only one patient was formally diagnosed in the medical record as being delirious at the time of the episode.

Chart review identified medication side-effects (48%, n = 30), fever (38%, n = 24), and infection (20%, n = 13) as the three most common problems noted in the two days prior to the onset of the first delirium episode.

The chart review indicated that a psychiatric consultation was ordered for approximately 37% (n = 51) of patients at some point during their stay. Multiple reasons for ordering a consultation were offered, but the most frequent were depressive symptoms (n = 20), mental status changes (n = 13), and medication side-effects/evaluation (n = 9). Analyses indicated that patients with pre-existing depression or dementia at admission were no more likely to generate a psychiatric consult request by staff than patients without depression or dementia (c 2 = 1.46, df = 1, p = .23 and c 2 = .147, df = 1, p = .70, respectively). Even delirious patients were no more likely to have a psychiatric consultation requested by staff than were non-delirious patients (33% vs. 41%; c 2 = 0.76, df = 1, p = .38). Although dementia may be difficult to distinguish from delirium retrospectively, our analysis indicated no significant relationship between dementia and delirium (c ² = 1.46, df = 1, p = .23).

The retrospective chart review, employing a diagnostic algorithm to determine the presence or absence of delirious symptoms in this SNF population, found that 46% of the patients had at least one, and in some cases, multiple episodes of delirium during their stay. This percentage is consistent with those reported in the literature for hospitalized medical admissions of persons with HIV/AIDS: prospective methods found that 30-57% of such patients had an episode of delirium.^6,7

The objective methods used in our study found that health care providers under-estimate the occurrence of delirium in AIDS patients. The diagnostic algorithm indicated delirious symptoms in 46% of the patients' charts, yet only one patient was diagnosed in the charts as being delirious. The fact that delirium could be retrospectively diagnosed from the patient charts indicates that staff were documenting symptoms consistent with delirium, but did not recognize when the symptom profile met criteria for a diagnosis of delirium. The problem is clear. When delirium is not recognized, it goes untreated when it can be managed. The retrospective chart review found delirium to be most often associated with fever, medication effects and infection. All have the potential of being alleviated, thus reducing the negative patient outcomes associated with delirium (e.g., decreased independence requiring institutionalization, increased length of hospitalization, death).^{9 - 14}

The chart review indicated that psychiatric consultations were ordered on a little more than one-third of the patients, even though 74% of the patients were documented as having some psychiatric morbidity. Depressed patients and patients with dementia were no more likely to have a psychiatric consult than patients without depression or dementia. One potential reason for this lack of relationship is that patients with pre-existing depression or dementia may have had ongoing psychiatric care outside of the SNF--in which case an internal psychiatric consultation would be less likely.

Similarly, delirious patients were no more likely to have a psychiatric consult than were non-delirious patients. However, the fact that the great majority of delirium episodes began after admission, and given the lack of relationship between delirium and depression or dementia, pre-existing psychiatric care is an unlikely explanation for the lack of an increase in psychiatric consultation for delirious patients. More likely is the explanation that providers' do not recognize the symptoms for a diagnosis of delirium. This lack of symptom recognition is exacerbated by barriers to recognizing the need for, and ordering a psychiatric consultation (e.g., physician opinion of need, other providers' lack of knowledge on how to determine the need for a psychiatric consultation).

Similar scenarios were indicated by the providers' learning needs assessment survey. Although three-quarters of the staff reported being comfortable ordering or advocating for a psychiatric consult, the majority of providers reported actually doing so for only a small percentage (0 - 10%) of their patients . This percentage was not proportionate to the percentage of patients (over half) they identified as being diagnosed with HIV/AIDS related neuropsychiatric conditions. Providers' estimates of the occurrence of delirium in their patients were extremely broad, ranging from 0% - 90%. The direct care providers' average estimate was 36% compared to 46% as assessed by the CAM. One might argue that two different time periods were used for the needs assessment (January - June, 1995) and the chart review (1994 patients) making comparisons between the two inappropriate. Review of the SNF's statistics, however, show that their provider staff and patients have not changed significantly over the two years of the study. Over 15 physicians, 40 nurses and aides, five social workers, a pastoral counselor as well as dietitians, occupational therapists and 230 volunteers care for patients at this SNF. A 0.1 full time equivalent psychiatrist is available to the patients, staff and families of patients. Staff turn-over is very low. For example, between 1994 and 1995, 38 of the SNF's 40 nurses remained constant. There was no change in administrative staff during the study period. Hence, the same providers completing the needs assessment survey cared for the 1994 patients.

Another finding related to patient length of stay would suggest that, if anything, providers' estimates of the occurrence of delirium, as measured on the 1995 needs assessment survey, should have been higher, not lower, than the occurrence estimates from the 1994 chart review. Between these two different time periods, patient demographics did not significantly change, but patient length of stay did. It increased from 45 to 63 days. Previous studies indicate that a bout of delirium is significantly associated with increased length of stay.^12,16

On the needs assessment survey, only 15% of the providers reported receiving less than 11 hours of HIV/AIDS education and training in the past three years. Approximately two-thirds, however, reported receiving 11 hours or less of neuropsychiatric education/training in the same time period. Given the data suggesting that psychiatric symptoms among HIV-infected populations are greater than in the community at large,^17-20 the paucity of neuropsychiatric training hours alone indicates a large need for neuropsychiatric training in general, and delirium training in particular, for providers taking care of terminally ill persons with AIDS.

Physician training, in particular, may eradicate or weaken several barriers to other providers (e.g., nurses, nurse practitioners, social workers) ordering or advocating for a psychiatric consult for their AIDS patients. A number of the staff indicated via the needs assessment survey that physicians attitudes about the need for and benefits of a psychiatric consult were significant barriers to ordering or advocating for psychiatric evaluation. Training and improved screening techniques may also alleviate the perceived barriers of lack of knowledge, questions regarding the necessity of psychiatric evaluation, and confusion regarding whose role it is to facilitate such an evaluation. Provider training would need to be accompanied by patient and/or family education on delirium, its management, and the benefits of a psychiatric consult to diagnose delirium in an effort to minimize stigma associated with psychiatric evaluation and patients’ refusal of such interventions. Barriers of cost and limited availability of psychiatric services could be overcome if future research clearly demonstrates the benefit to AIDS patients who receive psychiatric consultation services.

Comparing the objective and subjective findings of the study, it would seem that a self-report needs assessment survey which asks providers to rate their learning needs is reliable. Providers were frank about identifying their delirium learning needs; these were validated by the results of the chart review and providers' biographical data. On the other hand, the data suggest that one cannot rely on provider report of psychiatric morbidity alone in order to estimate the actual service needs of AIDS patients in a SNF.

In conclusion, the problem with delirium is that staff often do not recognize the symptoms as indicators of delirium, do not advocate for a psychiatric consult when patient signs and symptoms suggest the need, do not investigate the underlying reason(s) for the delirium, and do not work to alleviate these underlying problems. These combined self-report and chart review results highlight the need to train nurses, nurse practitioners, social workers, occupational therapists, physicians, and other providers involved in either the direct or indirect care of terminally ill AIDS patients, to appreciate the frequency of delirium in their patient population, to recognize the signs and symptoms of delirium, and to encourage staff to advocate for a psychiatric consult.

1. Trzepacz PT, Teague GB, Lipowski ZJ. Delirium and other organic mental disorders in a general hospital. Gen Hosp Psychiatry, 1985; 7:101-106.
2. Hale M, Koss N, Kertein M, et al. Psychiatric complications in surgical ICU. Crit Care Med, 1977; 5:199-203.
3. Blachy PH, Starr A. Post-cardiotomy delirium. Am J Psychiatry, 1964; 121:371-375.
4. Massie JF, Holland JC. Delirium in terminally ill patients with cancer. Am J Psychiatry, 1983; 140:1048-1050.
5. Johnson JC, Kerse NM, Gottlieb G, et al. Prospective versus retrospective methods of identifying patients with delirium. J Am Geriatr Soc, 1992; 40:316-319.
6. Fernandez F, Levy JK, Mansell PWA. Management of delirium in terminally ill AIDS patients. Intl J Psychiatry Med, 1989; 19:165-172.
7. Breitbart W, Marotta R, Platt M, Corbera K. Pharmacologic management of delirium in medically hospitalized AIDS patients, in Abstracts of the 37th Annual Meeting of the Academy of Psychosomatic Medicine. Chicago, APM, 1990.
8. Uldall KK, Koutsky L, Bradshaw D, Krone M. Utilization of hospital services by psychiatrically ill persons with AIDS: Washington State, 1990 - 1992. Unpublished manuscript.
9. Pasacreta JV, Massie MJ. Nurses' reports of psychiatric complications in patients with cancer. ONF, 1990; 17:347-353.
10. Levkoff SE, Evans DA, Liptzin B, et al. Delirium. Arch Intern Med, 1992; 152:334-340.
11. Francis J, Kapoor WN. Prognosis after hospital discharge of older medical patients with delirium. J Am Geriatr Soc, 1992; 40:601-606.
12. Thomas RI, Cameron DJ, Fahs MC. A prospective study of delirium and prolonged hospital stay. Arch Gen Psychiatry, 1988; 45:937-940.
13. Rabins PV, Folstein MF. Delirium and dementia: diagnostic criteria and fatality rates. Brit J Psychiatry, 1982; 140:149-153.
14. Wise MG. Reducing confusion in delirium diagnosis. Clin Psychiatric News, 1994; February: 2, 22.
15. Inouye SK, van Dyck CH, Alessi CA, Balkin S, Siegal AP, Horwitz RI. Clarifying confusion: The confusion assessment method. Annals Int Med, 1990;113:941-948.
16. Uldall, KK & Berghuis, JP. Delirium in AIDS patients: Recognition and medication factors. Under review for Am J Psychiatry.
17. Maj J. Organic mental disorders in HIV-1 infection. AIDS, 1990; 4:831-840.
18. Perry S, Jacobsberg LB, Fishman B, et al. Psychiatric diagnosis before serological testing for the human immunodeficiency virus. Am J Psychiatry, 1990; 147:89-93.
19. Tross S, Hirsch D, Rabkin B et al. "Determinants of current psychiatric disorder in AIDS spectrum patients." Program and abstracts of the Third International Conference on AIDS, June 1 - 5, 1987. Abstract T.10.5. Washington DC.
20. Atkinson JH, Grant I, Kennedy CJ et al. Prevalence of psychiatric disorders among men infected with human immunodeficiency virus: A controlled study. Arch Gen Psychiatry, 1988; 45:859-864.

Variable	Number and Rounded Percents
Most Frequent AIDS-associated Conditions
Candidiasis	27 (20%)
Cytomegalovirus (non-CNS)	48 (35%)
HIV encephalopathy (Dementia)	75 (55%)
Herpes simplex	21 (15%)
Kaposi's sarcoma	27 (20%)
Mycobacterium avium complex	52 (38%)
Neurological non-CNS	47 (34%)
Neurological - other (e.g. CVA, seizures)	28 (21%)
PCP	22 (16%)
Pneumonia - other	18 (13%)
Wasting syndrome	75 (55%)

Most Frequent Psychiatric Morbidity Conditions
Alcohol abuse/dependency	5 ( 4%)
Anxiety disorder	11 (8%)
Delirium	1 (<1%)
Dementia	75 (55%)
Insomnia	3 ( 2%)
Mood disorder	30 (22%)
Personality disorder	1 (<1%)
Psychotic disorder	2 ( 1%)
Substance abuse/dependence	8 ( 6%)

The Trouble With Delirium In AIDS Patients: The Discrepancy Between Occurrence And Health Care Provider Identification