People with intellectual disabilities (ID) represent a small but important group of hospitalized patients who have higher rates of complications than do patients without ID hospitalized for the same reasons. Complications are potentially avoidable conditions, such as healthcare-acquired infections, healthcare-acquired skin breakdown, falls, and medication errors and reactions. Addressing factors related to complications can focus efforts to improve hospital care. The purpose of this exploratory study was to analyze data from reviews of academic medical center charts (N = 70) about complications and to examine patient and hospitalization characteristics in relation to complications among adult patients (age ≥ 18 years) with ID hospitalized for nonpsychiatric reasons. Adults with ID tended to be twice as likely to have complications (χ2 = 2.893, df = 1, p = .09) if they had a surgical procedure and were nearly four times as likely to have complications (χ2 = 6.836, df = 1, p = .009) if they had multiple chronic health conditions (three of the following: history of cerebral palsy, autism spectrum symptoms, aggressive behavior, respiratory disorder, and admission through the emergency department). Findings suggest preliminary criteria for assessing risk for complications among hospitalized people with ID and the need for attention to their specific needs when hospitalized.
People with intellectual disabilities (ID) experience high rates of hospitalization. Compared to people without ID, those with ID are six times as likely to be hospitalized for ambulatory-sensitive conditions (Balogh, Brownell, Ouellette-Kuntz, & Colantonio, 2010). As well, a study in two regions in the State of New York among adults with ID age 40–79 years living in community residential facilities of 4–15 people indicated that 16% were hospitalized in 1 year (Janicki et al., 2002). In a systematic review of literature, it was found that, despite 20 years of efforts, people with intellectual disabilities continue to have poor hospital outcomes and experiences (Iacono, Bigby, Unsworth, Douglas, & Fitzpatrick, 2014). When hospitalized, people with ID had higher risks of complications after inpatient major surgery than did patients without ID; overall complications odds ratio was 1.53, compared to patients without ID (Lin, Liao, Chang, Chang, & Chen, 2011). Patients with ID also had higher rates of complications (unintended adverse events; UHC [formerly known as University Healthsystem Consortium], 2013) than did patients without ID hospitalized with the same diagnoses valid at discharge for three of the five main diagnoses for patients with ID (Ailey, Johnson, Fogg, & Friese, 2014).
Medicare Conditions of Participation are health and safety standards necessary for quality care and protecting the health and safety of patients; as part of the conditions of participation, hospitals are required to report complications (Office of Federal Registrar, 2014). Complications include postoperative complications, healthcare-associated infections, healthcare-acquired skin breakdown, medication errors and reactions, falls, blood transfusion reactions, and hospital-acquired deep vein thrombosis (Hughes et al., 2006; Levinson, 2010a). Complications do not always involve healthcare errors (Hughes et al., 2006).
In 2000, the Institute of Medicine (IOM) issued the report To Err Is Human: Building a Safer Health System (IOM, 2000). This landmark report highlighted safety problems and inadequate and faulty systems and processes that facilitate complications. The report focused attention on correcting serious quality control problems in health care in the United States. In 2003, the IOM released the report Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Disparities exist in the quality of health care that are not related to access to care but to characteristics of the populations (IOM, 2003), with much data on disparities faced by racial and ethnic minorities (Gold, 2014). Defining the disparities is difficult for some subgroups, including for people with disabilities, due to lack of data (Gold, 2014). People with intellectual disabilities, as other people with disabilities, likely face health disparities; and data regarding their healthcare are lacking. Addressing and reducing complications are key components of efforts to improve the safety and quality of healthcare. Despite ongoing efforts since the release of the IOM report in 2000, rates continue to be high; and disparities continue to exist (Gold, 2014; Levinson, 2010a). Among efforts to reduce complications is identifying risk factors associated with complications and taking steps to manage those risks (Healthgrades, 2013). People with multiple chronic conditions are shown to have poor health outcomes (Parekh, Goodman, Gordon & Koh, 2011), and complications when hospitalized are shown to be associated with comorbid chronic health conditions and demographics (Healthgrades, 2013). Addressing factors related to complications among patients with ID may help address the care needs of this population when hospitalized and may provide insights overall into efforts to reduce complications among other at-risk populations.
An exploratory retrospective review was conducted of data for people with ID (age ≥ 18 years) hospitalized during the period from January 1, 2008, to December 31, 2009, at a single academic medical center in the U.S. Midwest to describe patient and hospitalization characteristics potentially related to complications in this population. The Institutional Review Board of the medical center in the Midwest approved the retrospective chart review research with exempt status. Graduate nursing students who conducted the review all completed training in human subjects protection.
First, records from all patients with ID admitted during the 2-year period were identified using secondary diagnoses codes for ID. The following secondary International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9; National Center for Health Statistics, 2011) diagnosis codes were included in our definition of ID: 317 – mild mental retardation, 318.x – other specified mental retardation, 319 – unspecified mental retardation. All identifying information was stripped from the data. Information gathered was exclusively about patients admitted for medical or surgical reasons; psychiatric admissions were excluded. Following collection of data for all patients identified as having ID (N = 217), chart reviews were conducted on 70 cases, representing 32% of the sample.
Nurse or physician review of charts is identified as an effective method of determining complications (Levinson, 2010b). Complications were assessed in the review of 70 charts of adults with ID. Graduate nursing students conducted the reviews. The students participated in initial discussions regarding the information to be collected and in the development of a data collection form.
Because the study was exploratory, and previous research indicates that demographics and comorbid chronic health conditions can be related to complications, descriptive data were collected on several demographic variables including age, gender, length of stay, admission through the emergency department, and location before admission. Surgical procedures are related to complications in previous research among people with ID. Data were collected on surgical procedures and frequencies collected. Data also were collected on previous health history of chronic health conditions, and frequencies were calculated. Data were collected on complications, including healthcare-acquired infection, healthcare-acquired skin breakdown, medication errors/reactions, falls, postoperative complications, blood transfusion reactions, and hospital-acquired deep vein thrombosis; and frequency of complications was calculated.
Factors Associated With Complications
Two statistical tests were used in the analysis of factors associated with complications, the chi-square and binary logistic regression tests. The chi-square test is used to determine whether there is a significant difference between expected frequencies and observed frequencies. The chi-square test has the advantage that it makes no assumptions about the distribution of the population. Because a previous study indicated that surgical procedure was a risk factor for complications (Lin et al., 2011), a chi-square test was used to determine if surgical procedure was associated with complications in this study. The chi-square was set up with a two by two cross-tabulation table, with surgical procedure coded as 1 and not surgical procedure coded as zero, and complication coded as 1 and no complication coded as zero.
Because having multiple chronic conditions is associated with poor health outcomes, we conducted an exploratory backward binary logistic regression to determine possible factors related to complications. The use of a logistic regression with a sample of 70 is an exploratory analytic approach. With such a sample, a difference between proportions of .50 and .75 will be detected with a power of over .85, approximately the same for similar differences between proportions across most proportions. The detectable difference becomes small at the extreme proportions (less than .10 and greater than .90). Thus, this analytic approach can only be expected to detect relatively large effects, giving a general picture of factors that might be related to complications, without detail on how such factors are related. Because only a general picture of factors could potentially be determined with the small sample size, we did not attempt to determine a model that would predict complications. The backwards logistic regression was followed by a chi-square test to determine if the presence of multiple factors is related to complications.
For the logistic regression, the dependent variable, complication, was coded as a dichotomous variable, with any complication coded as 1 and no complications coded as zero. Independent variables entered into the equation were gender; age at admission; hospital length of stay; location prior to hospitalization; admission to the hospital through the emergency department (ED); arrival by ambulance; and previous history of seizures, cerebral palsy, gastrointestinal disease, respiratory disease, autism spectrum symptoms, and aggressive behavior. The follow-up chi-square was conducted with predictors with an odds ratio of 1.3 or above. The chi-square was set up as a two by two cross-tabulation with multiple factors (with the odds ratio of 1.3 or above) coded as 1 and no multiple factors coded as zero, and complication coded as 1 and no complication coded as zero.
Of the 70 patients whose charts were reviewed, 49% were males and 51% female; 55.7% were admitted to the medical center through the ED and 42.9% arrived at the medical center by ambulance. Before admission, 41.4% came from home, 24.3% from community-living facilities, 5.7% from nursing homes, and 25.7% from other hospitals. Frequencies were obtained on comorbid conditions listed in the previous health history. The most frequent previous health histories included history of seizures, gastrointestinal problems, cerebral palsy, respiratory conditions, autism spectrum disorders, and aggressive behaviors. Surgical procedures were noted for 19 of the 70 patients (27.1%; see Table 1). Of the 70 patients whose charts were reviewed, complications were noted for 16 (22.9%) and included healthcare-acquired infection, healthcare-acquired skin breakdown, medication errors/reactions, falls and postoperative complications. No cases with blood transfusion reactions or hospital-acquired deep vein thrombosis were found.
Factors Associated With Hospital Complications
Before calculating the chi-square tests and backwards logistic regression, we checked to determine if there was an association between particular diagnoses and complications; none was found. A chi-square test showed an association between any surgical procedure and complications. Although not reaching statistical significance in this small sample, patients with a surgical procedure as a risk factor were 2.1 times more likely to have any complication as those who did not have a surgical procedure (36.8% vs. 17.6%; χ2 = 2.893, df = 1, p = .09).
A backwards logistic regression was calculated. Surgical procedure was not entered into the logistic regression equation because it was close to statistically significant as an independent predictor of complications. At Step 6 of the backwards logistic regression, improvement in the prediction of complications was minimal. At this step, predictors with an odds ratio of 1.3 or above were retained for a follow-up chi-square analysis. The retained predictors were admission to the hospital through the ED, and history of cerebral palsy, autism spectrum symptoms, aggressive behaviors, and respiratory disorders (see Table 2). The range of risk factors was zero to 3. When subjects had three risk factors, they were found to be nearly four times as likely (60% compared to 17%) to experience a complication. This difference is statistically significant (χ2 = 6.836, df = 1, p = .009).
The finding on risk of complications postsurgical procedures is similar to the Lin et al. (2011) study, continuing to indicate surgical procedure as a risk factor for complications among patients with ID. Because multiple chronic conditions are known to be associated with poor health outcomes, the finding that the presence of multiple chronic conditions, along with admission through the emergency department, is related to complications during hospitalization is also a useful initial indicator of complication risk. It is not known whether the presence of multiple conditions or staff ability to manage, or a combination, is related.
Information about people with ID during hospitalization is often inadequate for planning comprehensive care (Bollands & Jones, 2002). The findings of this study are useful in providing preliminary guidance on information that should be gathered and used to plan care. Surgical procedures and the presence of multiple chronic conditions (history of cerebral palsy, respiratory disorder, aggressive behaviors, and autism spectrum symptoms), along with admission through the emergency department, are possibly associated with hospital complications among adult patients with ID. The presence of the multiple chronic conditions can be assessed for patients with ID at admission and used to classify a specific patient as being at high risk. It may be useful to specifically evaluate patients with ID, especially those with multiple chronic conditions, for heightened precautions to prevent complications. Heightened precautions may enable hospital staff to circumvent some of these complications.
In the general population, some events, such as falls and healthcare-associated infections, have been found to be associated with unmet nursing care needs (Lucero, Lake, & Aiken, 2010). Patients with ID may have difficulty communicating their needs to hospital staff (Iacono & Davis, 2003) and nurses may feel fearful and vulnerable when caring for people with ID (Sowney & Barr, 2006), possibly leading to unmet care needs. Specific standards of care for patients with ID as a way to reduce the incidence of hospital complications in this population may be useful. Addressing the care needs of this population also may provide lessons on addressing the needs of other populations that may be at increased risk of complications.
It should be noted that, with a sample of only 70, the findings of this study are preliminary. In addition, chart reviews involve the subjective judgment of the reviewer. The medical center may not be representative of other hospitals or even other academic medical centers. Of note is that nearly 25% of admissions of patients with ID came from other hospitals. As an academic medical center, this hospital may treat patients with ID difficult to treat at other hospitals. It is also possible that some patients are inappropriately transferred. As well, considering the lack of training of healthcare professionals in their care, it is possible that some individuals with ID did not have a secondary diagnosis indicating ID in their charts. Individuals without the secondary diagnosis noted would not be included in the database, and data may be incomplete.
In 2001, the Surgeon General of the U.S. Public Health Service issued the report Closing the Gap: A National Blueprint to Improve the Health of Persons With Mental Retardation (now called intellectual disability) that set a goal of improving the quality of health care for this population. The report noted the need for credible evidence-based standards of health care, based on scientific evidence for people with ID. This study provides preliminary evidence for improving standards of care for patients with ID. A larger study addressing factors related to complications in a predictive model, including factors such as age and whether admissions were for ambulatory sensitive conditions, would be useful.
The authors wish to acknowledge the funding support of the Center for Clinical Research and Scholarship at Rush University Medical Center and the Gamma Phi Chapter of Sigma Theta Tau.
Sarah H. Ailey, Department of Community, Systems and Mental Health Nursing; Tricia J. Johnson, Department of Health Systems Management, College of Health Sciences; Louis Fogg, Department of Community, Systems and Mental Health Nursing; and Tanya R. Friese, Department of Community, Systems and Mental Health Nursing, Rush University Medical Center, Chicago, Illinois.