Patient Safety Events and Harms During Medical and Surgical Hospitalizations for Persons With Serious Mental Illness
Abstract
Objective:
This study explored the risk of patient safety events and associated nonfatal physical harms and mortality in a cohort of persons with serious mental illness. This group experiences high rates of medical comorbidity and premature mortality and may be at high risk of adverse patient safety events.
Methods:
Medical record review was conducted for medical-surgical hospitalizations occurring during 1994–2004 in a community-based cohort of Maryland adults with serious mental illness. Individuals were eligible if they died within 30 days of a medical-surgical hospitalization and if they also had at least one prior medical-surgical hospitalization within five years of death. All admissions took place at Maryland general hospitals. A case-crossover analysis examined the relationships among patient safety events, physical harms, and elevated likelihood of death within 30 days of hospitalization.
Results:
A total of 790 hospitalizations among 253 adults were reviewed. The mean number of patient safety events per hospitalization was 5.8, and the rate of physical harms was 142 per 100 hospitalizations. The odds of physical harm were elevated in hospitalizations in which 22 of the 34 patient safety events occurred (p<.05), including medical events (odds ratio [OR]=1.5, 95% confidence interval [CI]=1.3–1.7) and procedure-related events (OR=1.6, CI=1.2–2.0). Adjusted odds of death within 30 days of hospitalization were elevated for individuals with any patient safety event, compared with those with no event (OR=3.7, CI=1.4–10.3).
Conclusions:
Patient safety events were positively associated with physical harm and 30-day mortality in nonpsychiatric hospitalizations for persons with serious mental illness.
Persons with serious mental illnesses, such as schizophrenia and bipolar disorder, have a mortality rate two to three times higher than the rate for the overall U.S. population, largely due to cardiovascular disease (1–6). Rates of other somatic conditions (for example, obesity, diabetes, hypertension, hyperlipidemia, and HIV/AIDS) (7–15) and health risk behaviors (for example, tobacco smoking, alcohol use, and physical inactivity) are also elevated in this group (9,16–21). Persons with serious mental illness are frequent consumers of nonpsychiatric health care services, and medical care that may not meet their unique needs could also play a role in poor health outcomes (22–25). Most previous studies examining quality of care for somatic conditions in this population have focused on established process-of-care metrics, such as prescription of beta-blockers after acute myocardial infarction (25). Despite the recent emphasis on patient safety and its role in preventable mortality in the overall U.S. population (26–30), little is known about patient safety in medical and surgical hospital care for persons with serious mental illness.
In the past 15 years, patient safety has risen to the forefront of efforts to improve health care delivery nationally. The 2000 Institute of Medicine (IOM) report To Err Is Human: Building a Safer Healthcare System (31), focused national attention on adverse events in the system and their role in morbidity, mortality, and health care spending. Adverse events are physical injuries incurred as a result of medical management, not as a result of underlying disease, and may be related to erroneous or delayed diagnostics, treatment errors, or failure to prevent injury (32). Research suggests that adverse events during hospitalization may affect as many as one in every ten patients (33) and that such events contribute to 14% to 50% of in-hospital deaths (34–36).
Despite considerable attention after release of the IOM report devoted to measuring and improving patient safety in the general U.S. population, vulnerable populations such as those with serious mental illness have infrequently been the focus of patient safety initiatives. This group may be at elevated risk of experiencing adverse events for a variety of reasons, including, but not limited to, potential interactions between psychotropic medications and commonly used analgesics or anesthetics, overuse of antipsychotics or other medications to address behavioral issues, communication challenges related to cognitive impairment and psychotic symptoms, and providers’ lack of experience caring for this group (32,37,38).
Three prior studies have considered patient safety in medical and surgical hospitalizations for persons with serious mental illness (32,37,38). These studies used patient safety indicators (PSIs), which are defined by using ICD-9 codes and other variables available in administrative claims data (39). All three studies compared PSIs among patients with and without serious mental illness and found that persons with serious mental illness were at elevated risk of experiencing many types of adverse events (32,37,38). Although these studies provided an important first step in understanding patient safety for this population, using adverse event indicators calculated from administrative data provides very limited detail on the events of interest. Compared with measurement tools that extract detailed information from medical records, PSIs have been shown to significantly underrepresent serious adverse events (29).
To our knowledge, no prior research has used medical record–based measures to characterize nonpsychiatric patient safety for persons with serious mental illness. This study addressed this gap and achieved four objectives: to describe the prevalence of patient safety events and related harms in medical-surgical hospitalizations for a community-based cohort of adults with serious mental illness; to explore the association between specific patient safety events and resultant physical harm; to examine the likelihood of death within 30 days if a patient safety event occurred; and to examine the association between any physical harm, severe physical harm, and increased length of stay and mortality.
Methods
Design and Population
We used a case-crossover design in which each participant served as his or her own control. For example, in our mortality analyses, participants’ hospitalizations that occurred within 30 days of death served as cases, and the same individuals’ hospitalizations that did not occur within 30 days of death served as controls.
The study sample was drawn from a parent cohort, which is described in detail elsewhere (3). Briefly, the parent cohort included Maryland Medicaid beneficiaries ages 21 to 64 with serious mental illness living in the Baltimore or Eastern Shore regions. Serious mental illness was defined as having any schizophrenia diagnosis, being disabled (defined as receiving Supplemental Security Income) with a diagnosis of bipolar disorder or major depression, or being disabled with another mental disorder diagnosis and use of specialty mental health care. For this study, we matched the parent cohort of Maryland Medicaid beneficiaries with serious mental illnesses to the National Death Index and selected the subset of individuals who died within 30 days of a medical-surgical hospitalization and who also had at least one prior medical-surgical hospitalization within five years of death. All hospitalizations occurred from 1994 to 2004.
We requested 1,601 hospital records for 425 adults with serious mental illness meeting these criteria. Of these, 790 records for 253 participants (60%) were received and reviewed, 476 records were not available from the hospital (for example, the record could not be located or was destroyed), and 335 records were received but not able to be reviewed because of insufficient data. Individuals whose charts were reviewed did not differ by age, sex, race, or psychiatric diagnoses from those whose charts were not reviewed.
Data
A multidisciplinary team of internists, patient safety experts, critical care physicians, nurses, and psychiatrists developed and tested the medical record abstraction tool, which emphasizes detection of patient safety events and related harms. To develop the tool, these stakeholders reviewed a small sample (74 charts of 47 patients) of medical-surgical hospitalizations of patients with serious mental illness at Johns Hopkins Hospital. All available chart data were reviewed, including physician, nurse, and consultant notes; discharge summaries; medication lists; medication dispensing logs; and any other included materials. This pilot review served as the basis for the development of patient safety event and harm categories. After reaching consensus on a set of categories, the group of experts worked together to develop a comprehensive list of consensus-based events and harms within each category. [A copy of the abstraction tool is available from the authors upon request.]
Data were abstracted by a team of five nurses and physicians. Interrater reliability was assessed to ensure coding consistency across reviewers. Cohen’s kappa statistics for all event and harm codes met or exceeded accepted standards (40). In addition to using the chart data, we used Maryland Hospital Association and Maryland Medicaid administrative claims data to measure hospital, hospitalization, and patient characteristics.
Measures
Key measures were patient safety events, harms, and mortality. Events were defined as situations with an untoward occurrence and grouped into eight categories: medication-related events, medical events, neurological-psychiatric events, hospital-acquired infections, unexpected procedure-related occurrences, care delivery events, unanticipated transfer to the intensive care unit, and unanticipated surgery or invasive procedure. Patient safety events were coded for clinical occurrences that were deemed “untoward occurrences’” not related to the primary reasons for admission—for example, pneumonia developed during hospitalization was coded, and pneumonia as the reason for admission was not. Consistent with prior patient safety assessment tools (30), events included occurrences, such as hospital-acquired infections, that are clearly iatrogenic in nature as well as other events that were unexpected and could signal human or system error but are less definitively iatrogenic (for example, myocardial infarction).
Harm was defined as a negative outcome directly related to the patient safety event. In the subset of the overall data used for this study, harms could be coded as physical harm or increased length of hospital stay due to the event. Severe harm was coded if the event was associated with permanent disability or an unexpected surgical procedure. Increased length of stay was defined as an additional one or more inpatient days caused by a patient safety event. Harms were coded only if the patient was directly affected by an event; near misses were not coded. Harms were always associated with events, but not all patient safety events had associated harms. An example of an event without a resultant harm is a scenario in which a patient received a different medication than prescribed without any detectable adverse effect. Mortality was measured with a dichotomous variable indicating hospitalizations within 30 days of death versus those not within 30 days of death.
Medicaid administrative claims data provided dates of hospitalization and patient age, sex, race, and psychiatric diagnoses. Hospitalization admission source and admitting service and patient marital status, living arrangements, and comorbid conditions were obtained from hospital charts. Chart-abstracted data on medical comorbidities were used to calculate the Charlson Comorbidity Index. Hospital teaching status and bed size were obtained from Maryland Hospital Association data.
Analysis
First, we examined patient characteristics at the time of first hospitalization and compared, using chi-square tests, patients’ comorbid conditions, admission source, admission service, length of stay, and hospital teaching status and bed size for hospitalizations within 30 days of death versus those not within 30 days of death. Second, we calculated the frequency and prevalence of patient safety events and related harms in medical and surgical hospitalizations for the community-based cohort of adults with serious mental illness in the five years preceding their death.
Third, we used conditional logistic regression to assess the relative likelihood of incurring any physical harm in hospitalizations with specific patient safety events versus those without such events. In this analysis, individual participants’ hospitalizations with and without physical harm served as cases and controls. The unit of analysis was event. Models were clustered at the hospitalization and person levels. As a result of this clustering, all hospitalization- and person-level characteristics were controlled.
Fourth, we used conditional logistic regression to compare the likelihood of death within 30 days of hospitalizations that included patient safety events versus those that did not include such events. In this analysis, each participant’s hospitalization within 30 days of death served as the case hospitalization, and the same individual’s hospitalization(s) that were not within 30 days of death served as the control(s). The unit of analysis was hospitalization. Models were clustered at the person level and controlled for admission source, admitting service, length of stay, and Charlson Comorbidity Index. Because of the case-crossover design, all time-invariant person-level characteristics (for example, age) were automatically controlled.
Fifth, we used conditional logistic regression to begin to examine the relationship between harm resulting from patient safety events and mortality. Although it was outside the scope of this study to perform detailed analysis of the relationship between the 11 specific types of harms measured in the abstraction tool and mortality within 30 days of hospitalization, we examined this relationship for three types of harm: any physical harm, severe physical harm, and increased length of stay. We used conditional logistic regression to formally test differences in the likelihood of these three categories of harms in hospitalizations that were within 30 days of death versus those that were not within this period. These models controlled for admission source, admitting service, length of stay, and Charlson Comorbidity Index and were clustered at the person level.
All study participants were decedents. The study was approved by the Johns Hopkins School of Medicine Institutional Review Board, and informed consent was waived.
Results
Schizophrenia was the most common primary psychiatric diagnosis in the study cohort (Table 1). Compared with hospitalizations not within 30 days of death, hospitalizations that occurred within 30 days of death had a higher associated Charlson Comorbidity Index (indicating more comorbidities) (4.3 versus 3.1, p<.001), a larger proportion involved a medical (versus surgical) admitting service, and length of stay was shorter. [A table summarizing these and other results of this analysis is available in the online supplement .]
| Characteristic | N | % |
|---|---|---|
| Age | ||
| <40 | 44 | 17 |
| 40–54 | 75 | 30 |
| 55–64 | 104 | 41 |
| ≥65 | 30 | 12 |
| Gender | ||
| Male | 156 | 62 |
| Female | 97 | 38 |
| Race | ||
| Black | 135 | 53 |
| White | 118 | 47 |
| Primary psychiatric diagnosis | ||
| Schizophrenia | 95 | 38 |
| Bipolar disorder | 49 | 19 |
| Major depression | 32 | 13 |
| Othera | 77 | 30 |
| Housing | ||
| Alone | 37 | 15 |
| With friends or family | 95 | 38 |
| With nonrelatives | 6 | 2 |
| Supported living | 52 | 21 |
| Jail or prison | 35 | 14 |
| Homeless | 5 | 2 |
| Other | 20 | 8 |
| Marital status | ||
| Single | 120 | 47 |
| Married | 28 | 11 |
| Separated, divorced, or widowed | 67 | 26 |
| Not specified or other | 35 | 14 |
a
For example, other psychosis, anxiety disorder, and personality disorder
We identified 4,547 patient safety events that occurred during 790 hospitalizations (Table 2), or a mean of 5.8 patient safety events per stay. The most frequently occurring categories of patient safety events were medication events (78% of the hospitalizations with safety events) and medical events (56% of the hospitalizations). Within these two categories, the most frequent events were prescribing errors (70%), dispensing errors (23%), electrolyte or acid-base disturbance (39%), and respiratory distress or failure (18%).
| Event | Hospitalizations with events (N=790) | Events (N=4,547) | |||
|---|---|---|---|---|---|
| Total N | Physical harm | ||||
| N | % | N | % | ||
| Medication | 614 | 78 | 1,876 | 111 | 6 |
| Prescribing | 550 | 70 | 1,496 | 28 | 2 |
| Dispensing | 183 | 23 | 271 | 12 | 4 |
| Reaction | 88 | 11 | 107 | 71 | 66 |
| Medical | 443 | 56 | 1,747 | 678 | 39 |
| Electrolyte or acid-base disturbance | 309 | 39 | 905 | 93 | 10 |
| Respiratory distress or failure | 139 | 18 | 210 | 141 | 67 |
| Integument event | 99 | 13 | 137 | 94 | 69 |
| Dysrhythmia | 88 | 11 | 133 | 67 | 50 |
| Gastrointestinal bleed or hemorrhage | 44 | 6 | 46 | 33 | 72 |
| Fall | 36 | 5 | 40 | 9 | 23 |
| Aspiration | 28 | 4 | 31 | 26 | 84 |
| Acute renal failure | 25 | 3 | 26 | 24 | 92 |
| Deep venous thrombosis or pulmonary embolus | 21 | 3 | 22 | 20 | 91 |
| Myocardial infarction | 18 | 2 | 18 | 15 | 83 |
| Neurologic or psychiatric | 182 | 23 | 317 | 182 | 57 |
| Mental status or level of consciousness | 131 | 17 | 208 | 134 | 64 |
| Psychiatric event | 33 | 4 | 39 | 4 | 10 |
| New seizures | 17 | 2 | 20 | 14 | 70 |
| Neurologic injury, transient ischemic attack, or stroke | 9 | 1 | 9 | 9 | 100 |
| Hospital-acquired infection | 76 | 10 | 104 | 86 | 83 |
| Hospital- or ventilator-acquired pneumonia | 42 | 5 | 43 | 35 | 81 |
| Central line–associated bloodstream infection | 21 | 3 | 23 | 15 | 65 |
| Procedure related | 151 | 19 | 237 | 94 | 40 |
| Surgical tube | 33 | 4 | 49 | 5 | 10 |
| Any tracheal tube | 30 | 4 | 44 | 31 | 71 |
| Central line | 25 | 3 | 28 | 12 | 43 |
| Operative or postoperative complication | 24 | 3 | 27 | 22 | 82 |
| Pneumothorax | 11 | 1 | 13 | 12 | 93 |
| Care delivery | 249 | 32 | 370 | 56 | 15 |
| Patient testing, overall | 162 | 21 | 193 | 5 | 3 |
| Delay in care delivery | 128 | 16 | 177 | 51 | 29 |
| Unanticipated transfer to intensive care unit | 90 | 11 | 105 | 52 | 50 |
| Unanticipated surgery or invasive procedure | 31 | 4 | 40 | 25 | 63 |
| Total | 727 | 92 | 4,547 | 1,121 | 25 |
| Total excluding medication events | 549 | 70 | 1,876 | 1,010 | 54 |
One quarter of patient safety events (25%) resulted in physical harm, with a rate of 142 physical harms per 100 admissions (data not shown). When medication events were excluded from the total, 54% of all other patient safety events resulted in physical harm. Events with the highest prevalence of physical harm were acute renal failure (92%), deep venous thrombosis or pulmonary embolus (91%), pneumothorax (93%), aspiration (84%), and myocardial infarction (83%).
Odds of physical harm were significantly elevated in hospitalizations during which 22 of the 34 measured patient safety events occurred (Table 3). Categories of events associated with elevated likelihood of harm included medical events (odds ratio [OR]=1.5), neurologic or psychiatric events (OR=2.3), hospital-acquired infections (OR=2.5), procedure-related events (OR=1.6), and unanticipated surgery or invasive procedure (OR=1.8). Odds of physical harm were lower in hospitalizations during which medication events occurred, compared with hospitalizations without such events (OR=.5).
| Event | OR | 95% CI |
|---|---|---|
| Medication | .5 | .4–.6 |
| Prescribing | .2 | .1–.3 |
| Dispensing | .3 | .2–.6 |
| Reaction | 2.8 | 2.1–3.6 |
| Medical | 1.5 | 1.3–1.7 |
| Electrolyte or acid-base disturbance | .3 | .2–.4 |
| Respiratory distress or failure | 2.4 | 1.9–2.9 |
| Integument event | 2.9 | 2.3–3.6 |
| Dysrhythmia | 1.9 | 1.4–2.5 |
| Gastrointestinal bleed or hemorrhage | 2.2 | 1.5–3.2 |
| Fall | 1.3 | .6–2.5 |
| Aspiration | 2.1 | 1.3–3.4 |
| Acute renal failure | 2.2 | 1.3–3.7 |
| Deep venous thrombosis or pulmonary embolus | 2.9 | 1.7–4.9 |
| Myocardial infarction | 2.1 | 1.3–3.4 |
| Neurologic or psychiatric | 2.3 | 1.9–2.8 |
| Mental status or level of consciousness | 2.4 | 2.0–2.9 |
| Psychiatric event | .6 | .2–2.0 |
| New seizures | 2.3 | 1.3–4.0 |
| Neurologic injury, transient ischemic attack, stroke | 4.0 | 1.8–8.9 |
| Hospital-acquired infection | 2.5 | 2.0–3.3 |
| Hospital- or ventilator-acquired pneumonia | 2.3 | 1.6–3.5 |
| Central line–associated bloodstream infection | 2.1 | 1.2–3.6 |
| Procedure related | 1.6 | 1.2–2.0 |
| Surgical tube | .3 | .1–.9 |
| Any tracheal tube | 1.8 | 1.2–2.8 |
| Central line | 1.7 | .9–3.2 |
| Operative or postoperative complication | 3.5 | 2.1–5.9 |
| Pneumothorax | 2.9 | 1.6–5.5 |
| Care delivery | .6 | .4–.8 |
| Patient testing, overall | .1 | .05–.4 |
| Delay in care delivery | 1.0 | .7–1.4 |
| Unanticipated transfer to intensive care unit | 1.3 | .9–1.8 |
| Unanticipated surgery or invasive procedure | 1.8 | 1.1–2.8 |
a
Reference group: hospitalizations without the indicated patient safety event. Relative odds were estimated by using conditional logistic regression models clustered at the hospitalization and person levels. As a result of clustering, all hospitalization- and person-level characteristics were controlled.
Patient safety events were more prevalent in hospitalizations that occurred within 30 days of death than in those not within 30 days of death (Table 4). Odds of death within 30 days of hospitalization were higher for hospitalizations that included any patient safety event (OR=3.7). When medication events—the most common type of event—were excluded from the total count of patient safety events, the likelihood of death within 30 days of hospitalization increased (OR=5.9). All categories of patient safety events except medication events, procedure-related events, and unanticipated surgery or invasive procedure were associated with increased likelihood of 30-day mortality.
| Event | Hospitalization | |||||
|---|---|---|---|---|---|---|
| Not within 30 days of death (N=488) | Within 30 days of death (N=302) | Relative odds of death | ||||
| N | % | N | % | ORa | 95% CI | |
| Medication | 390 | 80 | 224 | 74 | .7 | .5–1.0 |
| Prescribing | 356 | 73 | 194 | 64 | .7 | .5–.9 |
| Dispensing | 107 | 22 | 76 | 25 | 1.2 | .9–1.8 |
| Reaction | 43 | 9 | 45 | 15 | 1.6 | 1.0–2.6 |
| Medical | 201 | 41 | 242 | 80 | 6.7 | 4.5–10.1 |
| Electrolyte or acid-base disturbance | 142 | 29 | 167 | 55 | 3.8 | 2.6–5.5 |
| Respiratory distress or failure | 35 | 7 | 104 | 34 | 9.4 | 5.6–16.0 |
| Integument event | 36 | 7 | 63 | 21 | 3.5 | 2.1–5.7 |
| Dysrhythmia | 26 | 5 | 62 | 21 | 6.8 | 3.7–12.7 |
| Gastrointestinal bleed or hemorrhage | 10 | 2 | 34 | 11 | 6.4 | 2.9–14.1 |
| Fall | 28 | 6 | 8 | 3 | .3 | .1–.8 |
| Aspiration | 7 | 1 | 21 | 7 | 7.9 | 2.9–21.4 |
| Acute renal failure | 3 | 1 | 22 | 7 | 12.5 | 3.7–42.4 |
| Deep venous thrombosis or pulmonary embolus | 8 | 2 | 13 | 4 | 2.9 | 1.0–8.1 |
| Myocardial infarction | 3 | 1 | 15 | 5 | 8.0 | 2.2–28.5 |
| Neurologic or psychiatric | 76 | 16 | 106 | 35 | 3.4 | 2.3–5.1 |
| Mental status or level of consciousness | 47 | 10 | 84 | 28 | 3.9 | 2.6–5.9 |
| Psychiatric | 19 | 4 | 14 | 5 | 1.0 | .4–2.3 |
| New seizures | 4 | 1 | 13 | 4 | 5.3 | 1.7–16.9 |
| Neurologic injury, transient ischemic attack, or stroke | 0 | — | 9 | 3 | — | — |
| Hospital-acquired infection | 21 | 4 | 55 | 18 | 5.5 | 3.0–9.9 |
| Hospital- or ventilator-acquired pneumonia | 8 | 2 | 34 | 11 | 10.1 | 3.9–26.5 |
| Central line–associated bloodstream infection | 7 | 1 | 14 | 5 | 1.9 | 1.3–2.9 |
| Procedure related | 73 | 15 | 78 | 26 | 1.3 | .4–4.5 |
| Surgical tube | 11 | 2 | 22 | 7 | 2.2 | .9–5.4 |
| Any tracheal tube | 10 | 2 | 20 | 7 | 2.2 | .8–5.7 |
| Central line | 7 | 1 | 18 | 6 | 2.4 | .8–7.3 |
| Operative or postoperative complication | 13 | 3 | 11 | 4 | 2.7 | .6–12.9 |
| Pneumothorax | 3 | 1 | 8 | 3 | 2.2 | .5–10.0 |
| Care delivery | 128 | 26 | 121 | 40 | 2.0 | 1.2–3.2 |
| Patient testing, overall | 91 | 19 | 71 | 24 | 1.6 | 1.0–2.8 |
| Delay in care delivery | 56 | 12 | 72 | 24 | 1.9 | 1.1–3.3 |
| Unanticipated transfer to intensive care unit | 28 | 6 | 62 | 21 | 4.4 | 2.2–8.8 |
| Unanticipated surgery or invasive procedure | 15 | 3 | 16 | 5 | .8 | .3–2.2 |
| Total | 439 | 90 | 288 | 95 | 3.7 | 1.4–10.3 |
| Total excluding medication events | 286 | 59 | 263 | 87 | 5.9 | 3.2–10.7 |
a
Reference group: hospitalizations without the indicated patient safety event. Relative odds were estimated by using conditional logistic regression models clustered at the person level. Models controlled for admission source, admitting service, length of stay, and Charlson Comorbidity Index.
The prevalence of harm related to patient safety events was higher in hospitalizations that occurred within 30 days of death, compared with hospitalizations that did not occur with 30 days of death for any physical harm (56% versus 28%), severe physical harm (11% versus 3%), and increased length of hospital stay (42% versus 18%) (Table 5). Odds of death within 30 days of hospitalization were significantly elevated in hospitalizations with any of these three types of harm.
| Harm | Hospitalization | |||||
|---|---|---|---|---|---|---|
| Not within 30 days of death (N=488) | Within 30 days of death (N=302) | Relative odds of death | ||||
| N | % | N | % | ORa | 95% CI | |
| Any physical harm, excluding death | 137 | 28 | 168 | 56 | 3.3 | 2.3–4.9 |
| Severe physical harmb | 13 | 3 | 33 | 11 | 2.7 | 1.2–6.3 |
| Increased length of inpatient stayc | 88 | 18 | 126 | 42 | 2.8 | 1.8–4.2 |
a
Reference group: hospitalizations without a patient safety event. Relative odds were estimated by using conditional logistic regression models clustered at the person level. Models controlled for admission source, admitting service, length of stay, and Charlson Comorbidity Index.
b
The patient safety event was associated with permanent disability or an unexplained surgical procedure.
c
Increased length of stay was defined as 1 or more additional inpatient days as a result of a patient safety event.
Discussion
In our sample of persons with serious mental illness, patient safety events and associated harms were common in medical-surgical hospitalizations in the years preceding death. When the analyses adjusted for patient and hospitalization characteristics, we found that many of the patient safety events measured were positively associated with physical harm and that both events and harms were positively associated with 30-day mortality.
Three categories of patient safety events were associated with both physical harm and mortality, suggesting that these events may be important targets for future study and intervention: medical events, neurologic or psychiatric events, and hospital-acquired infections. Within these categories, patient safety events involving respiratory distress or failure, integumentary events, dysrhythmia, gastrointestinal bleed or hemorrhage, aspiration, acute renal failure, deep venous thrombosis or pulmonary embolus, myocardial infarction, changes in mental status or level of consciousness, hospital- or ventilator-acquired pneumonia, and central line–associated blood stream infections were all associated with increased likelihood of both physical harm and mortality within 30 days of hospitalization.
Although our study did not assess the causes of patient safety events and harms, prior research suggests several potential pathways unique to the population with serious mental illness. Medication interactions or overuse of psychotropic medications to address agitation or aggression could lead to changes in mental status, delirium, or oversedation, which could subsequently cause aspiration and respiratory complications (41,42). The counterintuitive finding that hospitalizations with medication events were less likely to involve any physical harm is likely due to the high frequency of these events and the fact that medication errors in and of themselves may not lead directly to physical harm. However, such events likely are direct contributors to other more harmful events—for example, to changes in consciousness and respiratory distress.
Sedation and use of restraints in response to behavioral issues may reduce mobility and increase risk of deep venous thrombosis or pulmonary embolus, respiratory failure, and hospital-acquired infection (37,43). Unfortunately, we were unable to determine from the chart data whether sedation and restraint use contributed to specific patient safety events. Underlying medical disease in the population with serious mental illness may also contribute to patient safety events and resultant harms; for example, chronic lung disease may contribute to respiratory-related events. In addition, medical and surgical providers’ inexperience caring for the special needs of persons with serious mental illness may contribute to negative patient safety outcomes. The active psychotic symptoms, cognitive impairment, and lack of social support that often accompany serious mental illness may lead providers to minimize or misinterpret somatic symptoms, potentially leading to harmful delays in diagnosis and treatment (44). Patient-provider communication challenges have been shown to be associated with suboptimal care in other vulnerable populations and may also play a role in adverse patient safety events among those with serious mental illness (45–48). Stigmatizing attitudes toward and beliefs about patients with serious mental illness among general medical providers may also play a role (49).
The proportion of medical-surgical hospitalizations in the study population with patient safety events and related harms—142 harms per 100 admissions—was considerably higher than rates in the general population. A study of ten North Carolina hospitals found a rate of 25 harms per 100 admissions (30), and another study of three large tertiary care centers found a rate of 49 harms per 100 admissions (29). However, these rates are not directly comparable with our rates because the studies used different patient safety measurement tools and our study population was limited to individuals with serious mental illness in the five years preceding their death. Because of our case-crossover design, we were unable to assess whether the apparent elevation in rates in our study sample compared with the general population was attributable to serious mental illness, elevated risk of adverse patient safety events during the last five years of life, or measurement differences.
Our study had several limitations. The generalizability of results is limited because only decedents were included in the sample. Our study population was limited to individuals with serious mental illness in the five years preceding their death. Estimates of the prevalence of patient safety events and harms reported in this study are therefore specific to this subset of Maryland’s population with serious mental illness and cannot be generalized to the overall population with serious mental illness in Maryland or the United States. Further, the lack of a comparison group without mental serious mental illness prevented us from assessing the relative rate of patient safety events and harms in populations with serious mental illness versus those without serious mental illness.
Despite these limitations, the case-crossover design provided the critically important benefit of automatically controlling for unmeasured individual-level risk factors (for example, tobacco smoking) that may influence patient safety, making this an optimal design to accurately assess the prevalence of patient safety events and related harms and the relationships among events, harms, and mortality among persons with serious mental illness. Although we controlled for key time-varying characteristics across hospitalizations, our analysis may have inadequately accounted for other time-variant factors. We used a newly developed measurement tool that has not been validated in other populations. However, the tool addresses key weaknesses present in other standard indicators, namely the PSIs and the global trigger tool (GTT). Unlike the PSIs, our tool was based on expert review of complete medical chart data. Unlike the GTT, our tool measured both events and harms and had clearly defined numerators and denominators for the measures of interest (28). We measured a wide range of patient safety events, defined as untoward and unexpected occurrences unrelated to the reason for hospital admission. Although some of these events are clearly iatrogenic in nature—for example, hospital acquired infections—it is possible that others may have occurred irrespective of human or system error. Completeness and accuracy of medical record abstraction may have been influenced by the legibility of handwritten notes. In addition, the content and format of medical notes may vary across individual providers and hospitals.
Conclusions
Nonpsychiatric hospitalizations are a vulnerable time for patients with serious mental illness. Patient safety in this population warrants additional attention in research and practice and better tools that allow health systems to consistently measure and track patient safety over time. Future research should assess patient safety in this population across diverse geographic regions and delivery settings and examine the unique system-, provider-, and patient-level factors that contribute to adverse outcomes among persons with serious mental illness. Many pathways to poor patient safety are likely shared by those with and without serious mental illness, and recent nationwide interventions, such as those shown to reduce central-line bloodstream infections, will likely benefit all patients (28). However, high rates of harmful events in this group suggest that those with serious mental illness may require special efforts. A better understanding of the modifiable causes of adverse events in this group and the development and testing of interventions to address those causes might reduce preventable mortality in this vulnerable population.
Footnote
The authors gratefully acknowledge support from grants R01MH074070 and K24MH093763 from the National Institute of Mental Health.
Supplementary Material
File (appi.ps.201500415.ds001.pdf)
- View/Download
- 53.00 KB
References
1.
Brown S: Excess mortality of schizophrenia: a meta-analysis. British Journal of Psychiatry 171:502–508, 1997
2.
Colton CW, Manderscheid RW: Congruencies in increased mortality rates, years of potential life lost, and causes of death among public mental health clients in eight states. Preventing Chronic Disease 3:A42, 2006
3.
Daumit GL, Anthony CB, Ford DE, et al: Pattern of mortality in a sample of Maryland residents with severe mental illness. Psychiatry Research 176:242–245, 2010
4.
Osby U, Correia N, Brandt L, et al: Mortality and causes of death in schizophrenia in Stockholm county, Sweden. Schizophrenia Research 45:21–28, 2000
5.
Parks J, Svendsen D, Singer P, et al: Morbidity and Mortality in People With Serious Mental Illness. Alexandria, Va, National Association of State Mental Health Program Directors, Medical Directors Council, 2006
6.
Saha S, Chant D, McGrath J: A systematic review of mortality in schizophrenia: is the differential mortality gap worsening over time? Archives of General Psychiatry 64:1123–1131, 2007
7.
Dickerson FB, Brown CH, Kreyenbuhl JA, et al: Obesity among individuals with serious mental illness. Acta Psychiatrica Scandinavica 113:306–313, 2006
8.
Allison DB, Newcomer JW, Dunn AL, et al: Obesity among those with mental disorders: a National Institute of Mental Health meeting report. American Journal of Preventive Medicine 36:341–350, 2009
9.
Daumit GL, Clark JM, Steinwachs DM, et al: Prevalence and correlates of obesity in a community sample of individuals with severe and persistent mental illness. Journal of Nervous and Mental Disease 191:799–805, 2003
10.
De Hert M, Van Winkel R, Van Eyck D, et al: Prevalence of diabetes, metabolic syndrome and metabolic abnormalities in schizophrenia over the course of the illness: a cross-sectional study. Clinical Practice and Epidemiology in Mental Health 2:14, 2006
11.
McEvoy JP, Meyer JM, Goff DC, et al: Prevalence of the metabolic syndrome in patients with schizophrenia: baseline results from the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) schizophrenia trial and comparison with national estimates from NHANES III. Schizophrenia Research 80:19–32, 2005
12.
Osborn DP, Levy G, Nazareth I, et al: Relative risk of cardiovascular and cancer mortality in people with severe mental illness from the United Kingdom’s General Practice Research Database. Archives of General Psychiatry 64:242–249, 2007
13.
Rosenberg SD, Goodman LA, Osher FC, et al: Prevalence of HIV, hepatitis B, and hepatitis C in people with severe mental illness. American Journal of Public Health 91:31–37, 2001
14.
Blank MB, Himelhoch SS, Balaji AB, et al: A multisite study of the prevalence of HIV with rapid testing in mental health settings. American Journal of Public Health 104:2377–2384, 2014
15.
Sajatovic M, Dawson NV: The emerging problem of diabetes in the seriously mentally ill. Psychiatria Danubina 22(suppl 1):S4–S5, 2010
16.
Casagrande SS, Anderson CA, Dalcin A, et al: Dietary intake of adults with serious mental illness. Psychiatric Rehabilitation Journal 35:137–140, 2011
17.
Compton MT, Daumit GL, Druss BG: Cigarette smoking and overweight/obesity among individuals with serious mental illnesses: a preventive perspective. Harvard Review of Psychiatry 14:212–222, 2006
18.
Leonard S, Adler LE, Benhammou K, et al: Smoking and mental illness. Pharmacology, Biochemistry, and Behavior 70:561–570, 2001
19.
LLerena A, de la Rubia A, Peñas-Lledó EM, et al: Schizophrenia and tobacco smoking in a Spanish psychiatric hospital. Schizophrenia Research 60:313–317, 2003
20.
Lohr JB, Flynn K: Smoking and schizophrenia. Schizophrenia Research 8:93–102, 1992
21.
Lopes FL, Nascimento I, Zin WA, et al: Smoking and psychiatric disorders: a comorbidity survey. Brazilian Journal of Medical and Biological Research 35:961–967, 2002
22.
Druss BG: Improving medical care for persons with serious mental illness: challenges and solutions. Journal of Clinical Psychiatry 68(suppl 4):40–44, 2007
23.
Druss BG: The mental health/primary care interface in the United States: history, structure, and context. General Hospital Psychiatry 24:197–202, 2002
24.
Druss BG, von Esenwein SA: Improving general medical care for persons with mental and addictive disorders: systematic review. General Hospital Psychiatry 28:145–153, 2006
25.
McGinty EE, Baller J, Azrin ST, et al: Quality of medical care for persons with serious mental illness: a comprehensive review. Schizophrenia Research 165:227–235, 2015
26.
Pronovost P, Holzmueller CG, Needham DM, et al: How will we know patients are safer? An organization-wide approach to measuring and improving safety. Critical Care Medicine 34:1988–1995, 2006
27.
Pronovost PJ, Goeschel CA, Colantuoni E, et al: Sustaining reductions in catheter related bloodstream infections in Michigan intensive care units: observational study. BMJ 340:c309, 2010
28.
Pronovost PJ, Wachter RM: Progress in patient safety a glass fuller than it seems. American Journal of Medical Quality 29:165–169, 2014
29.
Classen DC, Resar R, Griffin F, et al: “Global trigger tool” shows that adverse events in hospitals may be ten times greater than previously measured. Health Affairs 30:581–589, 2011
30.
Landrigan CP, Parry GJ, Bones CB, et al: Temporal trends in rates of patient harm resulting from medical care. New England Journal of Medicine 363:2124–2134, 2010
31.
Institute of Medicine Committee on the Quality of Health Care in America: To Err Is Human: Building a Safer Health Care System. Washington, DC, National Academy Press, 2000
32.
Daumit GL, Pronovost PJ, Anthony CB, et al: Adverse events during medical and surgical hospitalizations for persons with schizophrenia. Archives of General Psychiatry 63:267–272, 2006
33.
de Vries EN, Ramrattan MA, Smorenburg SM, et al: The incidence and nature of in-hospital adverse events: a systematic review. Quality and Safety in Health Care 17:216–223, 2008
34.
García-Martín M, Lardelli-Claret P, Bueno-Cavanillas A, et al: Proportion of hospital deaths associated with adverse events. Journal of Clinical Epidemiology 50:1319–1326, 1997
35.
García-Martín M, Lardelli-Claret P, Jiménez-Moleón JJ, et al: Proportion of hospital deaths potentially attributable to nosocomial infection. Infection Control and Hospital Epidemiology 22:708–714, 2001
36.
Dubois RW, Brook RH: Preventable deaths: who, how often, and why? Annals of Internal Medicine 109:582–589, 1988
37.
Khaykin E, Ford DE, Pronovost PJ, et al: National estimates of adverse events during nonpsychiatric hospitalizations for persons with schizophrenia. General Hospital Psychiatry 32:419–425, 2010
38.
Smith EG, Zhao S, Rosen AK: Using the patient safety indicators to detect potential safety events among US veterans with psychotic disorders: clinical and research implications. International Journal for Quality in Health Care 24:321–329, 2012
39.
Patient Safety Indicators: Software Documentation, Version 2.1, Revision 3A, SAS. Rockville, Md, Agency for Healthcare Research and Quality, 2003
40.
Landis JR, Koch GG: The measurement of observer agreement for categorical data. Biometrics 33:159–174, 1977
41.
Ely EW, Shintani A, Truman B, et al: Delirium as a predictor of mortality in mechanically ventilated patients in the intensive care unit. JAMA 291:1753–1762, 2004
42.
Copeland LA, Zeber JE, Pugh MJ, et al: Postoperative complications in the seriously mentally ill: a systematic review of the literature. Annals of Surgery 248:31–38, 2008
43.
Cooke BK, Magas LT, Virgo KS, et al: Appendectomy for appendicitis in patients with schizophrenia. American Journal of Surgery 193:41–48, 2007
44.
McDonald DD, Frakes M, Apostolidis B, et al: Effect of a psychiatric diagnosis on nursing care for nonpsychiatric problems. Research in Nursing and Health 26:225–232, 2003
45.
Cooper LA, Roter DL, Johnson RL, et al: Patient-centered communication, ratings of care, and concordance of patient and physician race. Annals of Internal Medicine 139:907–915, 2003
46.
Haywood C Jr, Bediako S, Lanzkron S, et al: An unequal burden: poor patient-provider communication and sickle cell disease. Patient Education and Counseling 96:159–164, 2014
47.
Lattimer L, Haywood C Jr, Lanzkron S, et al: Problematic hospital experiences among adult patients with sickle cell disease. Journal of Health Care for the Poor and Underserved 21:1114–1123, 2010
48.
Laws MB, Lee Y, Rogers WH, et al: Provider-patient communication about adherence to anti-retroviral regimens differs by patient race and ethnicity. AIDS and Behavior 18:1279–1287, 2014
49.
Mittal D, Corrigan P, Sherman MD, et al: Healthcare providers’ attitudes toward persons with schizophrenia. Psychiatric Rehabilitation Journal 37:297–303, 2014
Information & Authors
Information
Published In
Cover: Tea infuser and strainer, by Marianne Brandt, circa 1924. Silver and ebony. The Beatrice G. Warren and Leila W. Redstone Fund, 2000, The Metropolitan Museum of Art, New York City. Image copyright © The Metropolitan Museum of Art; image source: Art Resource, New York City.
History
Received: 24 September 2015
Revision received: 13 December 2015
Accepted: 1 February 2016
Published online: 16 May 2016
Published in print: October 01, 2016
Authors
Competing Interests
Dr. Pronovost reports receipt of grant support from Ernst and Young. The other authors report no financial relationships with commercial interests.
Funding Information
National Institute of Mental Health10.13039/100000025: K24MH093763, R01MH074070
Metrics & Citations
Metrics
Citations
Export Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.
View Options
View options
PDF/EPUB
View PDF/EPUBLogin options
Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.
Personal login Institutional Login Open Athens loginNot a subscriber?
PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.
Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).