Abstract
Background
Patient safety during the post-discharge period is a major public health concern. Racial differences on incidence and risk factors associated with post-discharge adverse events (AEs) are understudied. The aim of the study was to examine the differences on the incidence of post-discharge AEs and the associated risk factors between African American and Caucasian patients.
Methods
This was a prospective cohort study of patients at risk for post-discharge AEs from December 2011 through October 2012. We included 589 patients who were African American or Caucasian and discharged home from an urban community hospital. The patients spoke English and could be contacted after discharge for evaluation. Two nurses performed 30-day post-discharge telephone interviews, and two physicians adjudicated health records to determine AEs using a previously established methodology.
Results
African American patients had a slightly higher incidence of post-discharge AEs than Caucasian patients (30.6% vs. 29.9%), although the difference did not show statistical significance. The multivariable logistic regression model indicated that post-discharge AEs were associated with timely follow-up and the number of secondary discharge diagnoses. In subgroup analyses of the risk factors in each racial group separately, only timely follow-up ambulatory visits were associated with post-discharge AEs.
Conclusion
Post-discharge AEs were experienced by a large proportion of both African American and Caucasian patients, and there was no statistically significant difference in these proportions by race.
Keywords: Health Disparity, Transitional Care, Patient Safety, Race
INTRODUCTION
Post-discharge safety is a major public health concern for patients who are transferred from hospitals to home in the United States health care system.1 Recent studies found that the incidence of post-discharge AEs, which is defined as injury due to medical care in the 30 days after hospital discharge was 19%, or approximately five times higher than the estimated incidence of AEs during hospitalization.3 Most research on this subject tend to focus on the physical health outcomes of patients overall, rather than examining the unequal risk of post-discharge AEs between social and demographic groups.4 A better understanding of the incidence of post-discharge AEs and related risk factors in different socio-demographic groups may contribute to interventions to improve transitional care and reduce any inequities.
In the United States, racial disparities in the health care system and health outcomes exist even when insurance status, income, age, and severity of overall health conditions are comparable between African American and Caucasian patients.5 There are several reasons why these disparities occur in patient safety. African American patients tend to access low quality hospitals that are under resourced compared to Caucasian patients; these hospitals experience staffing shortages, lack of funding, and lack of technical support.6 Another explanation stems from preexisting discrimination towards African Americans, which leads to mistrust in medical professionals, depression, and impulsiveness.7,8 Finally, segregation and differences in socioeconomic status (SES) lead to African Americans experiencing difficulties accessing education and employment opportunities, which may further lead to poor health outcomes.9
Data related to racial differences on the incidence of post-discharge AEs and associated risk factors are understudied; and previous studies have been inconsistent with their results.4 One study found that African American and Caucasian patients had similar odds of having a readmission after discharge.10 Another found that African American patients were more likely to return to hospitals due to complications compared to Caucasian patients.11 However, another study did not find increased risk in readmission based on race.12 One study that was conducted in many hospitals across the country found that African American patients who were hospitalized were at a higher risk than Caucasian patients of experiencing nosocomial infections and adverse drug events.13 Data related to post-discharge safety needs to be analyzed further to identify commonalities and differences in risk factors for post-discharge AEs. In addition, African American patients experiencing AEs are considered a priority population by the Agency for Healthcare Research and Quality (AHRQ) because of their lower social economic status compared to Caucasian patients and the lack of research identifying evidence-based solutions to improve healthcare safety for African American patients.14
The objectives of the current study were to examine the racial differences on incidence of post-discharge AEs between African American and Caucasian patients who were discharged from an urban community hospital as well as identifying risk factors associated with post-discharge AEs. This study will help researchers better understand the racial differences in risk factors and develop specific interventions for African Americans to improve patient safety.
METHODS
Study Design
The Access Framework guided both our data collection and the analysis of the data.15 This study focused on three of the five components of the framework, specifically characteristics of the health care delivery system, utilization of health services, and characteristics of the population at risk. The remaining two components, health policy and consumer satisfaction, although important, were outside the scope of this particular work. The timing or timeliness of care falls under the utilization of the health services component of the model. The framework also incorporates elements of the population at risk that may be a contributing factor for adverse events. The framework enables us to examine the relationships between system and individual characteristics that may place individuals at greater risk for post-discharge AEs.
Study Setting, Population, and Recruitment
The study was conducted by an integrative team of hospitalist physicians, nurse-reviewers, physician-adjudicators, and researchers to determine the incidence and risk factors associated with post-discharge AEs. Eligible participants were recruited from the Tallahassee Memorial Hospital (TMH), an urban community hospital of Florida State University, from 14 December 2011, through 8 October 2012. Patients under the care of TMH physicians were recruited if they were being discharged home, spoke English, and could be contacted 30 days after discharge for a telephone interview.16 Potential participants were identified from a specially run list of pending discharges, limited to those patients on the general medical service and under the care of a TMH hospitalist physician. In order to insure that equal proportions of rural and urban subjects were recruited, the study employed an approach successfully utilized by an earlier study of post-discharge adverse drug events that sought to ensure adequate enrollment of racial/ethnic minorities.17 Nurse-reviewers initially stratified the patient list by rural/urban using zip codes, then randomized the order in which potentially eligible patients were approached to ensure unbiased patient enrollment.
Nurse-reviewers approached patients, at their bedside, to discuss the purpose of the study, invite their participation, and obtain informed written consent, including permission to obtain all health records from the patient’s outpatient physicians within 30 days of hospital discharge. A brief intake survey was then administered to collect potential confounders not easily obtained from administrative data sources, including socioeconomic factors (e.g., education level, household income), household (living alone or with a family member or friend), transportation (the ability to drive to and from the hospital), and caregiver status.
The study’s participation rate was 75% during recruitment for both African-American and Caucasian patients and the loss to follow-up rate was 4% (accounting for only 29 patients) for patients in both groups that we could not interview nor identify any health records within 30-days after hospital discharge.
Telephone Interviews
Participants were first contacted by nurse-reviewers by telephone within 3-4 weeks of discharge. If patients refused participation, they were excluded from the study. If the nurses were unable to contact the patients, attempts to gather post-discharge health records began to avoid loss to follow-up. TMH electronic health records were used to identify clinic visits, emergency department visits, and readmissions. Obituaries of local newspapers were also reviewed to identify any deceased patients.
The telephone interview consisted of three components: 1) Care Transition Measure-3 questions, which addressed the patients’ understanding of their health care needs and health management after discharge,18 2) questions regarding patients’ utilization of health services after discharge, including readmissions to the hospital, visits to the emergency department, visits to diagnostic facilities, planned or unplanned visits to their primary doctor’s office, or if they contacted their doctor’s office via telephone, and 3) questions addressing new or worsening symptoms since discharge that could represent post-discharge AEs.3,19-21 If any new or worsening symptoms were present, follow-up questions were asked to determine their relationship to medical care, if the patient was still able to perform daily activities, and the duration and severity of the symptoms.
Review of Health Records
Nurse-reviewers combined information obtained from the telephone interview and the outpatient health records to screen for: 1) new or worsening symptoms; 2) unplanned health services utilization; and 3) abnormal laboratory tests results. If nurse-reviewers identified any of the above information, they referred the case to physician adjudicators to be independently reviewed to determine the occurrence of post-discharge AEs. Two physician-adjudicators created case summaries for patients who may have experienced post-discharge AEs. The same physician-adjudicator would then rate their confidence that the patient injury was a result of medical management and not the patient’s underlying medical condition, including the absence of needed treatment when clearly clinically indicated (i.e., were post-discharge AEs), utilizing a 1 to 6 scale.3,16,17,19-23 If their rating was 4,5 or 6, the event was considered an AE. If there was any disagreement between the two physician-adjudicators, an independent third adjudicator would review the case and made a final decision.3,19
The determinations of AE severity were based on a combination of patient self-reported data during the phone interview and health record documentation by healthcare professionals.2 The physician adjudicators completed an AE assessment form for each AE which included examples of the severity categories, based on multiple prior studies.3,16,19 In general, a significant AE is one that causes self-limited signs or symptoms only (e.g., nausea, diarrhea, or rash). A serious AE is one that requires urgent or emergent treatment such as a hospitalization (e.g., gastrointestinal bleeding or delirium). A life-threatening AE is one that requires intensive care (e.g., respiratory failure requiring intubation). After identifying the study participants, a patients’ race was identified by computerized demographic data and arranged into two groups: African American and Caucasian. Of 603 total eligible participating patients, there were 589 who identified themselves either Caucasian or African American (465 were Caucasian, 124 were African American). Patients (N=14) who identified as other races (2.3%) were excluded from the current analysis.
Statistical Analysis
The proportion of African American and Caucasian patients with post-discharge AEs was determined. Health care utilization, functional consequences of post-discharge AEs, duration and severity of injury, symptoms associated with AEs, and drug categories implicated in adverse drug events were determined from the phone interview and medical record review. Bivariate analyses were first conducted to understand unadjusted relationships between various exposure variables and incidence of AEs in the entire study cohort. Chi-square tests (for categorical variables) and t-tests (for continuous variables) were utilized to examine differences in patient characteristics by race. Following bivariate analyses, multivariable logistic regression models were built to determine the independent factors associated with the odds of experiencing a post-discharge AE. Listwise deletion was used due to missing data.24 We included covariates that showed statistical significance in bivariate analysis in the final model as confounders. We also included patients’ age and gender in the multivariable logistic regression model as obligate covariates since they are both important demographic variables. Subgroup analyses were conducted to determine factors associated with AEs in African American and Caucasian patients separately; interaction terms were used to determine the presence of effect modification (i.e. whether certain factors had a different association with AEs by race). SPSS software (version 25) was used for all analyses.
RESULTS
Patient demographics are shown in Table 1. Slightly over half of the patients were female (n = 307, 52.1%), and the mean age of the 589 patients was 61.8 (S.D. = 15.0) with a statistically significant difference in age between African American (mean 55.5) and Caucasian (mean 63.4) patients (p < 0.001). Caucasian patients were generally more educated, had higher household incomes, had a higher proportion of patients being married, using Medicare, with planned clinic visits, and more often had a diagnosis of atrial fibrillation and coronary artery disease, whereas African American patients utilized mostly Medicaid and more often were diagnosed with hypertension.
Table 1.
Racial Differences in Demographic Characteristics and Clinical Diagnosis Among Patients Post-Discharge
| Characteristics | African American n (%) |
Caucasian n (%) |
p-value |
|---|---|---|---|
| Total number of sample (N = 589) | 124 | 465 | |
| Gender | .94 | ||
| Male | 59 (47.6) | 223 (48.0) | |
| Female | 65 (52.4) | 242 (52.0) | |
| Age, mean (SD) | 55.5 (14.4) | 63.4 (14.7) | <.001 |
| Education | .015 | ||
| Less than high school | 21 (16.9) | 42 (9.0) | |
| High school | 49 (39.5) | 151 (32.5) | |
| Some college | 26 (21.0) | 155 (33.3) | |
| College degree | 15 (12.1) | 65 (14.0) | |
| Post-graduate | 13 (10.5) | 52 (11.2) | |
| Household income* | .001 | ||
| < $25,000 | 56 (52.8) | 130 (31.9) | |
| $25,000 - $49,999 | 17 (16.0) | 107 (26.2) | |
| $50,000 - $74,999 | 17 (16.0) | 85 (20.8) | |
| $75,000 + | 16 (15.1) | 86 (21.1) | |
| Marital status | <0.001 | ||
| Married | 54 (43.5) | 296 (63.7) | |
| Divorced or Separated | 14 (11.3) | 45 (9.7) | |
| Widowed | 15 (12.1) | 70 (15.1) | |
| Never married | 41 (33.1) | 54 (11.6) | |
| Place of living | .78 | ||
| Urban | 62 (50.0) | 239 (51.4) | |
| Rural | 62 (50.0) | 226 (48.6) | |
| Living alone | .48 | ||
| No | 105 (84.7) | 381 (81.9) | |
| Yes | 19 (15.3) | 84 (18.1) | |
| Health insurance | <.001 | ||
| Private | 48 (38.7) | 165 (35.5) | |
| Medicare | 50 (40.3) | 263 (56.6) | |
| Medicaid | 19 (15.3) | 24 (5.2) | |
| Self-pay | 7 (5.6) | 13 (2.8) | |
| Number of secondary discharge diagnoses, mean (SD) | 10.9 (5.7) | 11.7 (5.2) | .12 |
| Number of days staying in hospital, mean (SD)* | 4.2 (3.4) | 3.7 (2.9) | .096 |
| Number of days after discharge for follow-up, mean (SD)* | 11.5 (18.2) | 10.3 (18.6) | .53 |
| Have a planned visit | 105 (84.7) | 428 (92.0) | .013 |
| Timely follow-up after discharge* | .25 | ||
| ≤ 7 days (timely follow-up) | 65 (52.8) | 270 (58.6) | |
| > 7 days (untimely follow-up) | 58 (47.2) | 191 (41.4) | |
| Primary discharge diagnosis | .10 | ||
| Circulatory | 39 (31.5) | 159 (34.2) | |
| Respiratory | 14 (11.3) | 56 (12.0) | |
| Gastrointestinal | 16 (12.9) | 79 (17.0) | |
| Genitourinary | 4 (3.2)^ | 26 (5.6) | |
| Infectious | 5 (4.0) | 34 (7.3) | |
| Neurovascular | 7 (5.6) | 14 (3.0) | |
| Other | 39 (31.5) | 97 (20.9) | |
| Top six diseases | |||
| Pneumonia | 10 (8.1) | 47 (10.1) | .49 |
| Hypertension | 100 (80.6) | 322 (69.2) | .012 |
| Type 2 diabetes mellitus | 48 (38.7) | 154 (33.1) | .24 |
| Atrial fibrillation | 10 (8.1) | 127 (27.3) | <.001 |
| Coronary artery disease | 27 (21.8) | 152 (32.7) | .019 |
| Infection | 22 (17.7) | 91 (19.6) | .65 |
Note:
Excludes missing cases; AE = Post-discharge adverse event
Count less than 5; Statistical significance: p < .05.
Among 177 patients who were identified with a post discharge adverse event, 38 were African American (30.6% of all African Americans) and 139 were Caucasian (29.9% of all Caucasians, Table 2). African American patients with post-discharge AEs were significantly more likely to report the experience of only symptoms (39.5%), while Caucasian patients generally were more likely to report the experience of mild effects on activities of daily living (39.9%) (p < 0.05). On the other hand, while functional consequences were less for African Americans, they suffered more severe AEs: 34.2% of African Americans suffered severe, life-threatening, or fatal AEs (i.e., worse than just “significant”), compared with 24.3% of Caucasians. Health care utilization, duration of injuries and symptoms associated with adverse events were similar between the two races. Visits to the ED occurred in 26% of African American patients and 22% of Caucasian patients, while readmissions to the hospital occurred in 11% of African Americans and 7% of Caucasians. The majority of patients with AEs experienced more than a week of symptoms.
Table 2.
Racial Differences in Patients’ Health Care Utilization, Functional Consequences, Duration and Severity of Injury, Symptoms, and Drug Classes Associated with Post-Discharge Adverse Events
| African American | Caucasian | p-value | |
|---|---|---|---|
| Total number of sample (N = 589) | 124 | 465 | |
| Subjects experiencing any AE | 38 (30.6%) | 139 (29.9%) | .87 |
| Outcome | n (% of patients with an AE) |
n (% of patients with an AE) |
|
| Health care utilization* | |||
| Readmission to hospital | 4 (10.5) | 10 (7.2) | .50 |
| Visit to ED | 10 (26.3) | 30 (21.6) | .54 |
| Visit to diagnostic facility | 1 (2.6) | 6 (4.3) | .64 |
| Planned visit to MD office | 16 (42.1) | 68 (48.9) | .46 |
| Unplanned visit to MD office | 2 (5.3) | 25 (18.0) | .053 |
| Telephone MD office | 1 (2.6) | 18 (12.9) | .069 |
| Functional consequences | .023 | ||
| Symptoms only | 15 (39.5) | 40 (29.0) | |
| Mild effects on ADLs | 11 (28.9) | 55 (39.9) | |
| Major effects on ADLs | 10 (26.3) | 43 (31.2) | |
| Death | 2 (5.3) | 0 (0.0) | |
| Duration of injury associated with AEs | .29 | ||
| Up to 1 day of symptoms | 1 (2.6) | 11 (7.9) | |
| 1 – 3 days of symptoms | 6 (15.8) | 13 9.4) | |
| 4 – 7 days of symptoms | 7 (18.4) | 16 (11.5) | |
| > 1 week of symptoms | 24 (63.2) | 99 (71.2) | |
| Severity of injury associated with AEs | .020 | ||
| Significant | 25 (65.8) | 103 (75.7) | |
| Serious | 11 (28.9) | 28 (20.6) | |
| Life-threatening | 0 (0.0) | 5 (3.7) | |
| Fatal | 2 (5.3) | 0 (0.0) | |
| Symptoms associated with AEs* | |||
| Gastrointestinal | 6 (15.8) | 28 (20.1) | .55 |
| Neuropsychiatric | 2 (5.3) | 22 (15.8) | .092 |
| Cardiovascular | 7 (18.4) | 17 (12.2) | .32 |
| Respiratory | 1 (2.6) | 5 (3.6) | .77 |
| Other | 6 (15.8) | 18 (12.9) | .65 |
| Drug categories implicated in adverse drug events | .76 | ||
| Cardiovascular | 5 (35.7) | 10 (26.3) | |
| Anti-infective | 2 (14.3) | 12 (31.6) | |
| Anti-coagulants | 1 (7.1) | 5 (13.2) | |
| Steroids | 1 (7.1) | 1 (2.6) | |
| Opioids | 1 (7.1) | 4 (10.5) | |
| Analgesics | 3 (21.4) | 4 (10.5) | |
| Diuretics | 1 (7.1) | 2 (5.3) | |
Note: AE = Post-discharge adverse event; ADLs = Activities of daily living; Statistical significance: p .05.
Health care utilization and symptoms associated with AEs are different than 100% because patients may have experienced multiple adverse events, a single adverse event may result in multiple types of care and symptoms, and there may be some missing data.
In Table 3, data from the full cohort were examined stratified by race to consider possible interactions between race and risk factors for AEs. Timely follow-up within seven days after discharge proved to be a significant positive predictor of post discharge AEs for both African American and Caucasian patients (p = .007 and p = .029, respectively). However, the association was much stronger for African-Americans (AE risk among those with timely follow up vs those without: 42% vs. 19%, rate ratio 2.21) than for Caucasians (34% vs. 25%, rate ratio 1.36).
Table 3.
Associations Between Patients’ Demographic Characteristics, Diagnosis and Experiences of Post-Discharge Adverse Events, by Race
| Characteristics | African American, n (% of AE) | Caucasian, n (% of AE) | ||||
|---|---|---|---|---|---|---|
| Have AE | No AE | p-value | Have AE | No AE | p-value | |
| Total number of sample (N = 589) | 38 | 86 | 139 | 326 | ||
| Gender | .975 | .120 | ||||
| Male | 18 (30.5) | 41 (69.5) | 59 (26.5) | 164 (73.5) | ||
| Female | 20 (30.8) | 45 (69.2) | 80 (33.1) | 162 (66.9) | ||
| Age, mean (SD) | 55.5 (14.4) | 55.5 (14.5) | .989 | 64.2 (14.9) | 63.1 (14.5) | .445 |
| Education | .135 | .438 | ||||
| Less than high school | 7 (33.3) | 14 (66.7) | 9 (21.4) | 33 (78.6) | ||
| High school | 14 (28.6) | 35 (71.4) | 51 (33.8) | 100 (66.2) | ||
| Some college | 4 (15.4)^ | 22 (84.6) | 44 (28.4) | 111 (71.6) | ||
| College degree | 8 (53.3) | 7 (46.7) | 22 (33.8) | 43 (66.2) | ||
| Post-graduate | 5 (38.5) | 8 (61.5) | 13 (25.0) | 39 (75.0) | ||
| Household income* | .631 | .517 | ||||
| < $25,000 | 17 (30.4) | 39 (69.6) | 44 (33.8) | 86 (66.2) | ||
| $25,000 - $49,999 | 4 (23.5)^ | 13 (76.5) | 27 (25.2) | 80 (74.8) | ||
| $50,000 - $74,999 | 6 (35.3) | 11 (64.7) | 27 (31.8) | 58 (68.2) | ||
| $75,000 + | 7 (43.8) | 9 (56.3) | 28 (32.6) | 58 (67.4) | ||
| Marital status | .380 | .900 | ||||
| Married | 19 (35.2) | 35 (64.8) | 85 (28.7) | 211 (71.3) | ||
| Divorced or separated | 6 (42.9) | 8 (57.1) | 14 (31.1) | 31 (68.9) | ||
| Widowed | 4 (26.7)^ | 11 (73.3) | 23 (32.9) | 47 (67.1) | ||
| Never married | 9 (22.0) | 32 (78.0) | 17 (31.5) | 37 (68.5) | ||
| Place of living | .436 | .752 | ||||
| Urban | 17 (27.4) | 45 (72.6) | 73 (30.5) | 166 (69.5) | ||
| Rural | 21 (33.9) | 41 (66.1) | 66 (29.2) | 160 (70.8) | ||
| Living alone | .324 | .619 | ||||
| No | 34 (32.4) | 71 (67.6) | 112 (29.4) | 269 (70.6) | ||
| Yes | 4 (21.1)^ | 15 (78.9) | 27 (32.1) | 57 (67.9) | ||
| Health insurance | .445 | .617 | ||||
| Private | 16 (33.3) | 32 (66.7) | 49 (29.7) | 116 (70.3) | ||
| Medicare | 16 (32.0) | 34 (68.0) | 82 (31.2) | 181 (68.8) | ||
| Medicaid | 3 (15.8)^ | 16 (84.2) | 6 (25.0) | 18 (75.0) | ||
| Self-pay | 3 (42.9)^ | 4 (57.1)^ | 2 (15.4)^ | 11 (84.6) | ||
| Number of secondary discharge diagnoses, mean (SD) | 12.1 (6.3) | 10.4 (5.4) | .131 | 12.4 (5.0) | 11.5 (5.3) | .089 |
| Number of days staying in hospital, mean (SD)* | 4.2 (4.3) | 4.2 (2.9) | .933 | 3.7 (2.1) | 3.7 (3.1) | .910 |
| Number of days after discharge for follow-up, mean (SD)* | 6.9 (5.4) | 13.6 (21.3) | .058 | 9.1 (8.8) | 10.9 (21.4) | .355 |
| Have a planned visit | 31 (29.5) | 74 (70.5) | .524 | 124 (29.0) | 304 (71.0) | .140 |
| Timely follow-up after discharge* | .007 | .029 | ||||
| ≤ 7 days (timely follow-up) | 27 (41.5) | 38 (58.5) | 92 (34.1) | 178 (65.9) | ||
| > 7 days (untimely follow-up) | 11 (19.0) | 47 (81.0) | 47 (24.6) | 144 (75.4) | ||
| Primary discharge diagnosis | .634 | .669 | ||||
| Circulatory | 11 (28.2) | 28 (71.8) | 53 (33.3) | 106 (66.7) | ||
| Respiratory | 3 (21.4)^ | 11 (78.6) | 18 (32.1) | 38 (67.9) | ||
| Gastrointestinal | 3 (18.8)^ | 13 (81.3) | 23 (29.1) | 56 (70.9) | ||
| Genitourinary | 2 (50.0)^ | 2 (50.0)^ | 7 (26.9) | 19 (73.1) | ||
| Infectious | 1 (20.0)^ | 4 (80.0)^ | 12 (35.3) | 22 (64.7) | ||
| Neurovascular | 3 (42.9)^ | 4 (57.1)^ | 4 (28.6)^ | 10 (71.4) | ||
| Other | 15 (38.5) | 24 (61.5) | 22 (22.7) | 75 (77.3) | ||
| Top six diseases | ||||||
| Pneumonia | 3 (30.0)^ | 7 (70.0) | .963 | 15 (31.9) | 32 (68.1) | .749 |
| Hypertension | 32 (32.0) | 68 (68.0) | .504 | 104 (32.3) | 218 (67.7) | .089 |
| Type 2 diabetes mellitus | 15 (31.3) | 33 (68.8) | .908 | 55 (35.7) | 99 (64.3) | .054 |
| Atrial fibrillation | 4 (40.0)^ | 6 (60.0) | .503 | 40 (31.5) | 87 (68.5) | .643 |
| Coronary artery disease | 10 (37.0) | 17 (63.0) | .415 | 47 (30.9) | 105 (69.1) | .736 |
| Infection | 10 (45.5) | 12 (54.5) | .097 | 31 (34.1) | 60 (65.9) | .332 |
Note:
Excludes missing; AE = Post-discharge adverse event
Count less than 5; Statistical significance: p < .05
Table 4 displays the multivariable logistic regression model for the entire cohort and the two racial subgroups before and after adjustment for potential confounders. In the overall cohort, race was not independently associated with post-discharge AEs (adjusted OR = 1.12, 95% CI .71 - 1.75, unadjusted odds ratio = .97). However, patients who did not have a timely follow-up (later than 7-days after discharge) had lower odds of experiencing a post-discharge AE than those who had a timely follow-up (aOR = .55, 95% CI .38 - .80, unadjusted odds ratio = .55). This association persisted in each of the subgroups (aOR = .30, 95% CI .13 - .70) in African Americans; (aOR = .64, 95% CI .42 - .97) in Caucasians. The effect size for timely follow up on AE risk was larger for the African-American patients compared to Caucasian patients (interaction p-value<0.05). Regarding other risk factors, for a one unit increase in a patient’s number of secondary discharge diagnoses, the adjusted odds of experiencing a post-discharge AE increased by approximately 3.7 percent (p = .034). Subgroup analyses by race showed similar effect sizes for number of diagnoses (but they were not statistically significant), with no significant interaction by race.
Table 4.
Multivariable Logistic Regression Model for Patients’ Risk of Experiencing a Post-Discharge Adverse Event
| All Patients | African American | Caucasian | ||||
|---|---|---|---|---|---|---|
| OR (95% CI) | aOR (95% CI) | OR (95% CI) | aOR (95% CI) | OR (95% CI) | aOR (95% CI) | |
| Race | ||||||
| African American vs. Caucasian | .97 (.63, 1.48) | 1.12 (.71, 1.75) | NA | NA | NA | NA |
| Gender | ||||||
| Male vs. Female | 1.29 (.90, 1.83) | 1.35 (.94, 1.94) | 1.01 (.47, 2.18) | 1.22 (.54, 2.77) | 1.37 (.92, 2.05) | 1.42 (.94, 2.13) |
| Age | 1.00 (.99, 1.02) | 1.00 (.99, 1.01) | 1.00 (.97, 1.03) | .99 (.96, 1.02) | 1.01 (.99, 1.02) | 1.00 (.99, 1.02) |
| Timely follow-up | ||||||
| Timely vs. Untimely | .55 (.38, .80) | .55 (.38, .80) | .33 (.15, .75) | .30 (.13, .70) | .63 (.42, .96) | .64 (.42, .97) |
| Number of secondary discharge diagnoses | 1.04 (1.00, 1.07) | 1.04 (1.00, 1.07) | 1.05 (.99, 1.12) | 1.07 (.99, 1.14) | 1.03 (1.00, 1.07) | 1.03 (.99, 1.07) |
Note: Dependent variable is the experiencing a post-discharge adverse event. Covariates: gender, age, timely follow-up, and number of secondary discharge diagnoses. OR = Unadjusted odds ratio; aOR = Adjusted odds ratio; CI = Confidence interval; Statistical significance: p < .05. NA = Not available.
DISCUSSION
In this prospective cohort study, we found just over 30% of patients suffered from a post-discharge AE, with an incidence almost identical in African American and Caucasian patients. There was a paradoxical result regarding the association between timely follow-up and post-discharge AEs, in which early follow-up care was associated with increased risk for AE. This association in our study was much larger for African-Americans than for Caucasians. The number of secondary diagnoses was also associated with AEs. There were no other identified risk factors and no other evidence of different risk factors or strength of associations between African American and Caucasian patients.
The paradoxical positive association between early follow-up and post-discharge AEs is consistent with a previous study conducted by our group, which showed the same association.25 Our findings may indicate that hospitalist physicians were accurately recognizing which patients were at a higher risk of developing an AE and therefore scheduled follow-up appointments sooner after discharge or took other measures to ensure timely follow-up.25 The risk assessments at discharge by clinicians may have been based in part on medical complexity (which we were able to adjust for, at least in part) but also psychosocial complexity, differences in health literacy, perceived stability at discharge, or functional status, which we were not able to adjust for (although we were able to account for living alone and insurance status). In our prior work, it was notable that the effect of early follow-up was attenuated with these adjustments, but did not disappear or reverse in direction, which might be expected if adjustment was complete. The fact that this association was greater among African Americans could indicate that they had a greater degree of psychosocial complexity and other unmeasured confounders. We cannot exclude the possibility that our findings also reflect information bias. Patients with faster follow-up may have had better documentation and patient awareness of symptoms that could be collected during the follow-up and outcome adjudication process. These findings illustrate the problems with determining the benefits of early follow-up in observational studies, where timing of follow-up is at the discretion of the medical team.
The number of secondary diagnoses (as a continuous variable) was associated with AE risk for the overall sample, with no evidence of an interaction by race. The lack of a statistically significant association in the subgroup analyses (with similar effect sizes and larger confidence intervals that just crossed unity) was likely due to smaller sample sizes in these cohorts.
The lack of other associations between patient characteristics and post-discharge AEs may reflect their relative lack of importance (compared with health system and provider characteristics and processes of care) or the importance of other characteristics that are harder to measure (such as social capital, degree of help from caregivers, health literacy, etc.). The association may have been somewhat weakened by the residential characteristics of the sample. Half of both the African-American and Caucasian patients in the study sample were from rural areas (Table 1). These patients are less likely to seek health care utilization, as shown by other studies,20,21 and therefore less likely to receive a prompt secondary diagnosis when compared to urban African-American and Caucasian patients who more frequently utilize the health care system.16
The lack of a difference in the rate of AEs by race was surprising given known disparities in health outcomes by race in many other contexts. Given the high rate of patient enrollment among both African-American and Caucasian patients (75%) and the low rate of loss to follow-up (4%) in our study, as well as expected differences in socioeconomic factors by race in the study cohort, we find it unlikely that this lack of effect can be explained by selection or loss-to-follow-up bias. On the other hand, we cannot rule out the effect of differential reporting by patients or documentation by their providers. These phenomena might also explain our finding that functional consequences were less frequent for African Americans, but that they suffered more severe AEs than Caucasians (i.e., that less severe AEs among African Americans were simply not reported by patients or elicited and documented by their providers). This could be due to the fact that African-American patients were less educated and perhaps less informed about their health conditions than Caucasian patients. It is also possible that compared to Caucasians, African Americans have worse functional status at baseline and/or generally under-report acute functional declines . Provider behavior may have also played a role: one previously conducted study found that physicians were more verbally dominant and tended to be less patient-centered in their approach with African American patients than with Caucasian patients,26 perhaps impacting their ability to detect post-discharge AEs, especially less severe events. Thus, it may be the case the post-discharge AEs are indeed more common among African Americans but that our approach to measurement, while rigorous, could not find it.
There were several limitations in the current study. During recruitment, patients were first stratified by urban and rural areas and enrollment was designed to ensure a sample equally distributed between rural and urban environments.16,17 Because of this study design, there was a smaller representation of African American patients compared to Caucasian patients as African-Americans were more likely to reside in urban areas. Another limitation is the lack of social determinants of health and other patient characteristics that might have served as risk factors or effect modifiers in our models. However, we were able to adjust for several clinical and sociodemographic variables. An additional limitation is the potential loss to follow-up bias in this study. Some patients were unavailable for the phone interview, but many of them had sufficient medical record data after discharge to allow for determination of AEs. Lastly, this study was conducted in one hospital system, limiting the potential generalizability of our findings.
CONCLUSION
Future studies need to focus more extensively on the differences in risk factors associated with post-discharge AEs among African American and Caucasian patients. There are previous studies that focus on AEs but neglect to discuss the differences between risk factors based on socio-demographic characteristics. Most patient safety research does not provide statistical analysis of the socio-demographic characteristics of patients experiencing post-discharge AEs. Increasing the sample sizes of African American and Caucasian patients in primary data collection, especially across multiple institutions, and capturing additional information on social determinants of health, will help researchers understand stronger associations among socio-demographic and AEs to develop trends and patterns in patient outcomes.4 By conducting more research on this topic, we will be able to develop interventions that will prevent or reduce AEs for patients with different racial backgrounds. This is important since racial disparities in the health care system exist, and we as researchers need to understand these disparities to reduce the incidence of AEs and the racial gaps in health care systems. Overall, more effective implementation of interventions is needed to improve patient safety for both African American and Caucasian patients during transitional care.
Acknowledgments
Funding: This study was funded by the Agency for Healthcare Research and Quality (grant R01HS018694 to Dr. Tsilimingras)
Grant Support: This work is supported by an R01 award from the Agency for Healthcare Research and Quality (grant R01HS018694 to Dr. Tsilimingras).
Footnotes
Conflict of Interest:
Mr. Costello declares that he has no conflict of interest.
Dr. Zhang declares that she has no conflict of interest.
Dr. Schnipper is the recipient of a grant from Sanofi-Aventis for an investigator-initiated study of intensive discharge interventions to improve outcomes in patients with diabetes discharged on insulin.
Dr. Tsilimingras declares that he has no conflict of interest.
Ethical approval: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Ethical approval: This article does not contain any studies with animals performed by any of the authors.
Informed consent: Informed consent was obtained from all individual participants included in the study.
Publisher's Disclaimer: Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality.
Publisher's Disclaimer: This Author Accepted Manuscript is a PDF file of a an unedited peer-reviewed manuscript that has been accepted for publication but has not been copyedited or corrected. The official version of record that is published in the journal is kept up to date and so may therefore differ from this version.
REFERENCES
- 1.Tsilimingras D, Bates DW. Addressing postdischarge adverse events: a neglected area. Joint Commission journal on quality and patient safety. 2008;34(2):85–97. [DOI] [PubMed] [Google Scholar]
- 2.Brennan TA, Leape LL, Laird NM, et al. Incidence of Adverse Events and Negligence in Hospitalized Patients. New England Journal of Medicine. 1991;324(6):370–376. [DOI] [PubMed] [Google Scholar]
- 3.Forster AJ, Murff HJ, Peterson JF, Gandhi TK, Bates DW. The incidence and severity of adverse events affecting patients after discharge from the hospital. Annals of internal medicine. 2003; 138(3): 161–167. [DOI] [PubMed] [Google Scholar]
- 4.Dahya V, Bishop-Royse J. The Sociodemographic Characteristics of Patients Who Experience Postdischarge Adverse Events. J Geriatr Med Gerontol. 2018;4(2). [Google Scholar]
- 5.Nelson A Unequal treatment: confronting racial and ethnic disparities in health care. Journal of the National Medical Association. 2002;94(8):666–668. [PMC free article] [PubMed] [Google Scholar]
- 6.Hasnain-Wynia R Disparities in health care are driven by where minority patients seek care: examination of the hospital quality alliance measures. Archives of internal medicine (1960). 167(12): 1233–1239. [DOI] [PubMed] [Google Scholar]
- 7.McClendon J, Bogdan R, Jackson JJ, Oltmanns TF. Mechanisms of Black-White disparities in health among older adults: Examining discrimination and personality. J Health Psychol. 2019:1359105319860180. [DOI] [PubMed] [Google Scholar]
- 8.Dovidio JF, Penner LA, Albrecht TL, Norton WE, Gaertner SL, Shelton JN. Disparities and distrust: the implications of psychological processes for understanding racial disparities in health and health care. Social science & medicine (1982). 2008;67(3):478–486. [DOI] [PubMed] [Google Scholar]
- 9.Williams DR, Collins C. Racial Residential Segregation: A Fundamental Cause of Racial Disparities in Health. Public Health Reports. 2001;116(5):404–416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pines JM, Mongelluzzo J, Hilton JA, et al. Postdischarge adverse events for 1-day hospital admissions in older adults admitted from the emergency department. Annals of emergency medicine. 2010;56(3):253–257. [DOI] [PubMed] [Google Scholar]
- 11.Silverstein MD, Qin H, Mercer SQ, Fong J, Haydar Z. Risk factors for 30-day hospital readmission in patients >/=65 years of age. Proceedings (Baylor University Medical Center). 2008;21(4):363–372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kassin MT, Owen RM, Perez SD, et al. Risk factors for 30-day hospital readmission among general surgery patients. Journal of the American College of Surgeons. 2012;215(3):322–330. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Coffey RM, Andrews RM, Moy E. Racial, ethnic, and socioeconomic disparities in estimates of AHRQ patient safety indicators. Medical care. 2005;43(3 Suppl):I48–57. [DOI] [PubMed] [Google Scholar]
- 14.Priority Populations. 2019: https://www.ahrq.gov/topics/priority-populations/index.html. [Google Scholar]
- 15.Aday LA, Andersen R. A framework for the study of access to medical care. Health Serv Res. 1974. Fall;9(3):208–220. [PMC free article] [PubMed] [Google Scholar]
- 16.Tsilimingras D, Schnipper J, Duke A, et al. Post-Discharge Adverse Events Among Urban and Rural Patients of an Urban Community Hospital: A Prospective Cohort Study. Journal of general internal medicine. 2015;30(8):1164–1171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Schnipper JL, Roumie CL, Cawthon C, et al. Rationale and Design of the Pharmacist Intervention for Low Literacy in Cardiovascular Disease (PILL-CVD) Study. Circulation: Cardiovascular Quality and Outcomes. 2010;3(2):212–219. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Parry C, Mahoney E, Chalmers SA, Coleman EA. Assessing the quality of transitional care: further applications of the care transitions measure. Medical care. 2008;46(3):317–322. [DOI] [PubMed] [Google Scholar]
- 19.Forster AJ, Clark HD, Menard A, et al. Adverse events among medical patients after discharge from hospital. CMAJ: Canadian Medical Association journal = journal de l'Association medicale canadienne. 2004;170(3):345–349. [PMC free article] [PubMed] [Google Scholar]
- 20.Schnipper JL, Kirwin JL, Cotugno MC, et al. Role of pharmacist counseling in preventing adverse drug events after hospitalization. Archives of internal medicine. 2006;166(5):565–571. [DOI] [PubMed] [Google Scholar]
- 21.Kripalani S, Roumie CL, Dalal AK, et al. Effect of a pharmacist intervention on clinically important medication errors after hospital discharge: a randomized trial. Annals of internal medicine. 2012; 157(1):1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Leape LL, Brennan TA, Laird N, et al. The nature of adverse events in hospitalized patients. Results of the Harvard Medical Practice Study II. The New England journal of medicine. 1991;324(6):377–384. [DOI] [PubMed] [Google Scholar]
- 23.Thomas EJ, Studdert DM, Burstin HR, et al. Incidence and types of adverse events and negligent care in Utah and Colorado. Medical care. 2000;38(3):261–271. [DOI] [PubMed] [Google Scholar]
- 24.Fox-Wasylyshyn SM, El-Masri MM. Handling missing data in self-report measures. Res Nurs Health. 2005;28(6):488–495. [DOI] [PubMed] [Google Scholar]
- 25.Tsilimingras D, Ghosh S, Duke A, Zhang L, Carretta H, Schnipper J. The association of post-discharge adverse events with timely follow-up visits after hospital discharge. PloS one. 2017;12(8):e0182669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Johnson RL, Roter D, Powe NR, Cooper LA. Patient Race/Ethnicity and Quality of Patient–Physician Communication During Medical Visits. American Journal of Public Health. 2004;94:2084–2090. [DOI] [PMC free article] [PubMed] [Google Scholar]