New tool only modestly better than hospice eligibility guidelines at predicting survival in advanced dementia.

Institution:

Icahn School of Medicine at Mount Sinai

Article Citation

PubMed ID:

Clinical Bottom Lines

In this study, the ADEPT (Advanced Dementia Prognostic Tool) was successfully validated in a cohort of 660 nursing home residents in Boston, Massachusetts. The sensitivity and specificity of the ADEPT score in predicting six-month mortality were 55.0% (95% CI, 45.2-64.4) and 71.3% (95% CI, 67.1-75.3) respectively. The sensitivity and specificity of hospice eligibility criteria were 20.0% (95% CI 13.0-28.0) and 89.0% (95% CI, 86.0-92.0) respectively. The ADEPT score only modestly improved prognostication as compared to the hospice eligibility requirements, which suggests that treatment decisions should be based on patient preferences rather than calculated life expectancy.

Disclaimer:

This is a review of the validity of a single study; the ‘bottom lines’ do not reflect comparison with the rest of the literature on this subject.

Methods

Type of Study:

Clinical Prediction Rule

Follow-up Period:

six months

Setting:

21 nursing homes in Boston, Massachusetts

Patient Population:

Among 1425 screened residents, 830 residents met eligibility criteria and 606 nursing home residents were recruited to participate in the study. Proxy refusal was the sole reason for nonrecruitment and nonparticipants were more likely to be men and less likely to be white. The mean age of participants was 85-89 and ranged from 65 to >100. 50.7% were aged 80-90 years. 18.2% were men and 88.0% were white. 42.2% of residents were completely functionally dependent and 9.7% were completely bedbound. 5.9% experienced dyspnea, 17.7% had congestive heart failure, 41.6% had insufficient oral intake, 8.3% had a BMI less than 18.5, and 11.7% had recent weight loss. 88.6% had bowel incontinence and 5.4% had at least one stage 2+ pressure ulcer.

Inclusion Criteria: age over 65, advanced dementia (as defined by Cognitive Performance Score of 5 or 6 and cognitive impairment due to any kind of dementia), presence of a health care proxy that could be identified and contacted

Significant Exclusions:

Resident’s whose proxy could not be identified or contacted were excluded. Those with cognitive impairment not due to dementia were excluded.

Intervention/Exposure:

Research assistants collected data from residents’ charts and nurse interviews to obtain demographic data and also to calculate each resident’s ADEPT score. The 12-item ADEPT score is a tool to predict six-month mortality in nursing home patients. It was initially derived using limited retrospective data from the MDS (Minimum Data Set) and was then re-derived based on a national MDS data set. The following elements are included in the ADEPT score: nursing home stay <90 days, age, male sex, shortness of breath, one or more pressure ulcers at stage 2+, activity of daily living score of 28 (i.e. complete functional dependence), bedfast most of day, insufficient oral intake, bowel incontinence, BMI less than 18.5, recent weight loss, congestive heart failure.

Data collected by research assistants from charts and nurses and health care proxy interiews was used to calculate whether or not residents met Medicare hospice eligibility for dementia. Residents with dementia are eligible for hospice if they have a score of 7c on the FAST (a dementia rating scale) as well as one medical condition in the last year including aspiration pneumonia, pyelonephritis, septicemia, multiple decubitus ulcers, recurrent fever after antibiotics or poor nutritional status.

Participants were then followed and it was determined whether or not they died within 6 months of their baseline evaluation.

Outcome Measures:

Mortality by six months.

Results

The mean ADEPT score of residents was 10.1 (SD 3.1.) Observed and predicted 6-month mortality were calculated and the goodness-of-fit test showed no significant lack of fit (P=0.69.) The AUROC (Area Under the Receiver Operating Characteristic Curve) for the ADEPT score’s prediction of six month mortality was 0.67 (95% CI 0.62-0.72.) This number changed only minimally when the cutoffs for the length of stay item were varied from the 90-day cutoff established by the study protocol. The operating characteristics of the ADEPT risk score were calculated at various cut points. A cut point of >11.0 achieved the highest AUROC: 0.63 (95% CI, 0.58-0.68) with a sensitivity of 55.0% (95% CI, 45.2-64.4) and specificity of 71.3% (95% CI, 67.1-75.3.) However, the ADEPT tool is a continuous measure and clinicians may select different cutoffs with different operating characteristics.

35.5% of residents were at a FAST 7c and 22.3% had one of the medical conditions within the last 12 months that were required for hospice eligibility. In all, 10.7% of residents were both FAST 7c and had >1 medical condition and therefore met hospice eligibility requirements. The AUROC for hospice eligibility guidelines was 0.55 (95% CI 0.51-0.59.) Sensitivity was 20.0% (95% CI 0.13-0.28) and specificity was 89.0% (95% CI, 0.86-0.92.)

When the cutoff score for ADEPT was set to 13.5, which also had the same specificity as the hospice eligibility criteria, the AUROC of the ADEPT score was 0.58 (95% CI, 0.54-0.63) and the sensitivity was 27.0% (95%CI, 19.0-36.0), which were not significantly different from that of the hospice eligibility guidelines.

Methodological Issues and Applicability To Older Adults

The cohort studied in this trial was predominately white and lived in the Boston area, which may limit generalizability.

Though the ADEPT score with a cutoff of 11.0 is modestly better than the hospice eligibility criteria at predicting six-month mortality, neither measure reliably predicts mortality in those with advanced dementia. This underscores the difficulties of prognostication even in those with clearly advanced disease. Considering the high burden of symptoms in advanced dementia and the poor prognosis overall, the delivery of palliative care should be guided by patient preferences rather than by the six-month prognosis requirement currently used by hospice.

Other Information

Created By:

Jennifer Reckrey, MD, Geriatric Fellow, Department of Geriatric and Palliative Medicine, Mt. Sinai School of Medicine

CAP Topics:

Dementia

Mortality

Comments

<p> It would be interesting and possible very relevant clinically if the ADEPT could be programmed into the MDS 3.0 so that anyone with a score of 11 or greater would be flagged. These would be particularly important patients to have a detailed goals of care and prognosis conversation with family. </p>

Search form

Comments