Outcomes for Hospitalized Older Adults with Down Syndrome in the United States

Data and sample

Retrospective data were pooled from the 2002 to 2012 Nationwide Inpatient Sample (NIS) databases (renamed as the National Inpatient Sample in 2012), sponsored by the Agency for Healthcare Research and Quality (AHRQ) as part of the Healthcare Cost and Utilization Project (HCUP). NIS is the largest publicly available all-payer inpatient care database in the United States, representing a 20% stratified sample of all US community hospitals [10]. A weighting variable provided by HCUP for trends and discharges was used to generate a national estimate [11]. Weighted frequencies are approximately five times the unweighted frequencies, because the NIS contains a 20% random sample from all hospitalizations in the US every year.

All patients aged 65 years or older who had an International Classification of Disease, 9th Revision, Clinical Modification diagnosis code indicating Down syndrome (ICD-9-CM: 758.0) were identified. Patients transferred to another short-term hospital were excluded in order to avoid double counting. The final study sample consisted of 2,134 records (unweighted sample), representing 10,144 discharges in the US (nationally weighted sample) from 2002 through 2012.

Outcomes

Outcome measures included in-hospital mortality and hospitalization associated costs. Costs were converted from the total charge variable in the NIS datasets using charge-to-cost ratios and were reported in 2014 US dollars using Consumer Price Index for hospital services.

The HCUP NIS contains data on total charges for each hospital discharge. This charge information represents the amount that hospitals billed for services, but does not reflect how much hospital services actually cost or the specific amounts that hospitals received in payment. The HCUP Cost-to-Charge Ratio Files are used to convert hospital charges to actual costs. For each data year, a Cost-to-Charge Ratio File was created based on all-payer inpatient cost for every hospital in the corresponding NIS database. Using the merged data elements from the Cost-to-Charge Ratio Files and the total charges reported in the NIS databases, we converted the hospital total charge data to cost estimates by multiplying total charges with the appropriate cost-to-charge ratio. Because the study sample contains data across 11 years, we reported costs data in 2014 US dollars using Consumer Price Index for hospital services [10].

Independent variables

Outcomes were adjusted for patient demographics (age, sex, and race/ethnicity), admission type (admitted from a long-term care facility and admission year), comorbidities, and hospital characteristics (size, rurality, teaching status, and region). Patients not admitted from long-term care facility may have been admitted from outpatient facilities, other acute care hospitals, or emergency departments. Age (in years) was transformed into four categories: 65–69, 70–74, 75–79, and 80 or older. Race/ethnicity was re-categorized into white, black, other race/ethnicity, and missing. Because some states do not report data on race/ethnicity, these data were not missing at random. Those discharges with missing race/ethnicity were assigned to a separate “missing” group. Admission year was grouped into three-year periods, except for the last two years. Comorbidities that are common for older patients with DS were included to adjust for patient risk factors. Hospital characteristics were obtained from the NIS hospital weights file.

Data analyses

Patient and hospital characteristics are presented as relative frequencies and percentages for each independent variable. T-statistics were calculated to analyze the overall trend of each independent variable during the study period. Trend p values were obtained using analysis of variance (ANOVA) with Bonferroni adjustments to analyze the differences among study variables across the 11 years. Univariate comparison between survivors and non-survivors was performed using t-test and Chi-square test. Multivariate logistic regression was conducted for in-hospital mortality. A generalized linear model with logit link and gamma distribution was fitted for right skewed cost data as a function of all the independent variables.

All statistical analyses were performed using SAS 9.4^® (SAS Institute, Inc., Cary, NC, 2013). All tests of significance were two-sided; significant results were defined at p < 0.05. Approval to conduct the study as exempt from human subjects review was received from the Institutional Review Board at the University of Arkansas for Medical Sciences.

Hospitalizations trends

During the 11-year study period, the number of hospitalizations with a diagnosis of DS among older adults (≥65 at the time of admission) showed an upward trend, with an increase from 649 hospital discharges in 2002 to 1,300 hospital discharges in 2012 (Fig. 1A). This represents the weighted frequency (labeled on y-axis left hand side). The weighted frequency, represents an adjusted value to account for the entire US population, and is derived from the un-weighted frequency weighted frequency (labeled on -axis right hand side). This is in contrast to a relatively stable trend for individuals with DS across all age groups (Fig. 1B).

OPEN IN VIEWER

Fig.1

A) Number of hospitalizations with a diagnosis of Down syndrome for older adults (≥65), 2002–2012. B) Number of hospitalizations with a diagnosis of Down syndrome for all age groups, 2002–2012.

The three most common primary diagnoses for hospitalization among older adults with DS were aspiration pneumonitis, acute pneumonia, and septicemia. These diagnoses accounted for 40.6% of all hospitalizations (Table 1). Operations on the digestive system, operations on the cardiovascular system, and diagnostic and therapeutic procedures were the most common procedures performed and accounted for 74.0% of all procedures (Table 1).

Patient and hospital characteristics

Table 2 shows the patient and hospital characteristics of the hospitalizations for older adults with DS. Patients 65–69 years of age accounted for 69% of the overall study population and 11% were older than 75 years of age at the time of admission. Majority of patients with DS with Alzheimer’s disease (76.6%) hospitalized were in the 65–69 years age group compared to 67.7% for patients with DS without Alzheimer’s disease (p value 0.009) (Supplementary Tables 1 and 2). More than half of the older patients with DS received care in large hospitals and 79.3% were treated at urban hospitals (including teaching and non-teaching).

A significant increase (based on trend p values) in hospitalization was observed in Southern states over the study period (from 19.7% in 2002 to 34.2% in 2012). The Midwest had a significant downward trend, from 30% in 2002 to 21.5% in 2012. Among comorbid conditions for older adults with DS, an overall upward trend was observed for hypothyroidism, mental disorders (including dementia), and hypertension, diabetes, and ear disorders. Hypothyroidism was the most common comorbidity (40%) among hospitalization diagnoses for DS patients.

Hospital mortality and costs

Table 3 depicts the outcomes of hospitalized older patents with DS. Hospital mortality rates and discharge disposition of older patients with DS to skilled nursing facilities have decreased significantly over the 11-year period while home discharges increased. However, no significant difference in temporal trends in cost of hospitalization was noted over the study period. The cost of hospitalization (overall and for top 5 primary diagnoses for hospitalization) for patients with DS with Alzheimer’s disease was significantly lower compared to patients with DS without Alzheimer’s disease (Supplementary Table 3).

Table 4 depicts the risk factors associated with hospital mortality and costs of hospitalization among older adults with DS. Significant risk factors associated with increased mortality during hospitalization included age 70–79 years at the time of admission [AOR1.42 (1.20–1.68) for age 70–74 years] and AOR 1.40 (1.07–1.84) for age range 75–79 years], female gender [AOR 1.20 (1.04–1.38)], patients in Western US region [AOR 1.45 (1.17–1.80)]; presence of ischemic heart disease [AOR 1.96 (1.59–2.42)], Alzheimer’s disease [AOR 1.52 (1.25–1.85)], cerebrovascular disease [AOR 1.73 (1.33–2.24)], and earlier era hospitalization (2002–2004). In contrast, presence of hypothyroidism, hypertension, eye and ear diseases were associated with lower mortality during hospitalizations. Among hospital characteristics, patients hospitalized to large urban hospitals compared to rural hospital [AOR 0.74 (0.62–0.89) and 0.56 (0.46–0.68) for urban non-teaching and urban teaching hospitals, respectively], in the Midwest region [AOR 0.79 (0.64–0.98)] had lower mortality rates.

Risk factors significantly associated with hospitalization costs included black race, medium and large hospitals, location in Western region, presence of comorbid chronic obstructive pulmonary diseases (COPD), and admission between 2005 and 2007. Meanwhile, female gender, Midwest and South locations, the presence of Alzheimer’s disease and cognitive disorders, and comorbid eye diseases were associated with lower hospitalization costs. Specifically, the mean cost of hospitalization for patients with DS and Alzheimer’s disease was $15,972.6 (SD: $38,272.2) per hospitalization compared to $18,621.85 (SD: $58,486.1) for patients with DS without AD. Supplementary Table 3 provides details of average cost of hospitalization for older patients with DS and AD compared to those without AD.

Supplementary Table 4 provides a univariable comparison of survivors and non-survivors. Care in urban teaching hospitals, and co-morbidities such as hypertension and hypothyroidism were more frequently noted in survivors compared to non-survivors, whereas stroke and ischemic heart disease was noted more frequently in non-survivors compared to survivors. However, after Bonferroni adjustments none of the comparisons were statistically significant.

Limitations

There are several limitations in the current study. First, the NIS database is an administrative database. Like most electronic databases, NIS lacks detailed information on laboratory, pathology, imaging, and patient severity (like degree of individual’s cognitive capacity). These missing confounders might affect outcome comparisons. The diagnosis of Alzheimer’s disease in individuals is not standardized and thus subject to variations. Finally, the NIS database is compiled from hospital discharge data and does not have any outpatient follow-up data for mortality beyond the hospitalization. The NIS is an administrative database that aims to gather data for billing purposes and can be limited by erroneous coding. However, the Healthcare Cost and Utilization Project quality control measures should minimize these possibilities. Furthermore, the hard clinical end-points used in our analysis are more difficult to miscode. Nevertheless, the potential for unmeasured confounders that may bias the outcomes still exist.

In summary, despite increasing life expectancy, and growing numbers of hospitalized elderly Americans with DS, there is a trend of decreasing hospital mortality and discharge disposition to skilled nursing facilities among individuals with DS, from 2002 through 2012. The present study highlights the continuing need for allied healthcare workers to be aware of the complex medical needs of older patients with DS. It is plausible that care goal discussions earlier on in treatment may reduce hospitalizations among some of these patients. Future studies will be of interest.