Abstract
Objective
To determine the predictive value of discharge destination as a surrogate for defining unfavorable outcome at 3- and 12-months poststroke.
Design
Analysis of the prospectively collected data from a randomized, placebo-controlled trial in patients with ischemic stroke presenting within 3 hours of symptom onset.
Setting
Post hoc analysis of patients recruited in a clinical trial.
Participants
Patients (N=530) discharged alive from the hospital after ischemic stroke.
Interventions
Not applicable.
Main Outcome Measures
Positive and negative predictive value and likelihood ratios of discharge destination for unfavorable outcome at 3- and 12-months poststroke defined by a Modified Rankin Scale (MRS) score of 2 to 6, 3 to 6, or 4 to 6. A likelihood ratio indicates how many times more (or less) likely a particular discharge destination is seen in patients with an unfavorable outcome compared with those without unfavorable outcome.
Results
The positive predictive value of nursing home and rehabilitation facility discharges was highest for unfavorable outcome defined by an MRS score of 2 to 6 (95%) and rehabilitation facility (89%) at 3-months poststroke, respectively. The positive predictive value of rehabilitation facility/nursing home (90%) was also highest for unfavorable outcomes defined by an MRS score of 2 to 6 compared with those defined by MRS scores of 3 to 6 (79%) and 4 to 6 (57%). The positive likelihood ratio was highest for nursing home discharges (13; 95% confidence interval [CI], 4.1– 41) followed by rehabilitation facility discharges for unfavorable outcome defined by an MRS score of 2 to 6 at 3-months poststroke (5.3; 95% CI, 3.5–7.9). The negative likelihood ratio was the highest for home discharge for unfavorable outcome defined by an MRS score of 2 to 6 (4.5; 95% CI, 3.4 – 6.1). A similar pattern was observed with unfavorable outcome defined using various thresholds at 12 months.
Conclusions
Discharge destination can provide high predictive values and likelihood ratios for death and disability at 3-months poststroke, as defined by an MRS of score of 2 to 6.
Keywords: Nursing homes, Patient discharge, Rehabilitation, Stroke
Ascertainment of disability at 3-months poststroke is an essential component of outcome assessment in stroke patients. The Modified Rankin Scale (MRS) has become one of the most widely used assessment tools to determine the severity of disability, because of its high interobserver reliability, superiority to other indices (eg, Barthel Index),1 and consistent use in trials and registries conducted in patients with stroke.1–6 However, large studies using datasets such as the Nationwide Inpatient Sample,7,8 National Hospital Discharge Survey,9 University HealthSystems Consortium,10 and Statewide Inpatient Sample11 do not have outcomes ascertained at 3-months poststroke using the MRS and have used discharge destination as a surrogate for defining favorable and unfavorable outcomes. Given the value of studies conducted using data from large administrative datasets, a better understanding of the relationship between discharge destination and MRS score at 3-months poststroke will allow a more accurate interpretation of results. We evaluated the predictive values of using discharge destination as the surrogate marker for 3 and 12 month MRS score using data collected as part of the randomized National Institute of Neurological Disorders and Stroke (NINDS) recombinant tissue plasminogen activator (rt-PA) Stroke Trial.12
METHODS
We used the public access data files available from the National Technical Information Services (Springfield, Virginia) for the NINDS rt-PA stroke trial that was a double-blinded, placebo-controlled, randomized trial that recruited patients from January 1991 through October 1994. The details of the trial have been previously published.12–14 The trial was conducted in 8 centers each of which developed a flowchart of acute stroke patient screening, assessment, and treatment and multidisciplinary teams that participated in treatment and hospital care in the preparatory phase of the trial.15 The study was approved by the Data Safety Monitoring Board appointed by NINDS and the local institutional review board at each of the participating sites. Detailed information on inclusion and exclusion criteria has been published in a previous publication.12 Briefly, eligible patients had an ischemic stroke with a clearly defined time of onset, a deficit measurable on the National Institutes of Health Stroke Scale score (NIHSS), and no evidence of intracranial hemorrhage on a prerandomization computed tomographic scan of the brain. After the neurologic deficits were quantified using the NIHSS, 624 patients were randomized to receive either placebo or rt-PA, 0.9mg/kg body weight within 3 hours of stroke onset.15
Disability and functional status was assessed using the Barthel Index, MRS, and Glasgow Outcome Scale and assessed at 3 and 12 months after randomization in all the randomized patients. The outcome was determined at 24 hours and 3-months poststroke by certified examiners who had not performed the baseline examination and was not present during the initial treatment and thus were unaware of the initial NIHSS.12 Certified nurse coordinators or study physicians determined the vital status and patient’s ability to perform daily activities (measured with the Barthel Index) and degree of functional disability (measured with the MRS) at 12 months using a telephone interview.13 The evaluators, patients, and their caregivers were unaware of the treatment assignments. The relatively homogenous cohort of ischemic stroke patients with well-defined and validated methods for outcome ascertainment up to 1 year provided adequate data for our analysis.
Statistical Analysis
We provided the data distribution (frequency for categorical and means for continuous variables) for baseline demographic and clinical variables according to discharge destination. The baseline variables assessed included age, sex, ethnicity/race, NIHSS strata (0 –9, 10 –19, or ≥20), hypertension, diabetes mellitus, angina pectoris, congestive heart failure, atrial fibrillation, hyperlipidemia, previous history of stroke or transient ischemic attack, and current cigarette smoking. The 4 discharge destinations (home, relative’s/friend’s home, rehabilitation facility, and nursing home) were analyzed individually. The outcome of interest was unfavorable outcomes defined by the MRS score at 3 and 12 months. The MRS score categorizes the magnitude of death or disability into 6 grades: 0 (no symptoms), 1 (no significant disability), 2 (slight disability), 3 (moderate disability), 4 (moderately severe disability), 5 (severe disability), and 6 (dead).16 We focused on the MRS score to define unfavorable and favorable outcomes rather than the Barthel Index because of increasing use of the MRS in clinical trials related to stroke. The preferential use of the MRS in trials is related to high interobserver reliability,16 smaller sample size requirement if the MRS score is used as a primary endpoint,17 and lower variation in methodology of using the MRS score as a primary endpoint in previous stroke trials.18
We sought to evaluate the positive and negative predictive value of various discharge destinations for 3 different definitions of unfavorable outcome at 3 and 12 months, defined by an MRS score of 2 to 6, 3 to 6, or 4 to 6. These 3 thresholds for defining unfavorable outcome have been used in previous clinical studies depending on the magnitude of benefit sought with intervention. Unfavorable outcome has been defined by an MRS score of 2 to 6 in trials evaluating the benefit of intravenous rt-PA,12,19 an MRS score of 3 to 6 in trials evaluating the benefit of endovascular treatment,20,21 and an MRS score of 4 to 6 in trials evaluating the benefit of hemicraniectomy.22 We evaluated the predictive value of discharge destination for all 3 definitions of unfavorable outcomes to provide a comprehensive analysis that can be used in a variety of settings for ischemic stroke patients. Unfavorable outcome (rather than favorable outcome) was chosen as the outcome of interest based on previous recommendations.18
The positive predictive value is the proportion of patients with a particular discharge destination who have experienced an unfavorable outcome. The negative predictive value is the proportion of patients discharged to a particular discharge destination who have experienced a favorable outcome. We also calculated the likelihood ratio, which provides a direct estimate of the odds of experiencing an unfavorable outcome based on discharge destination, either presence (positive) or absence (negative). A likelihood ratio indicates how many times more (or less) likely a particular discharge destination is seen in patients with an unfavorable outcome compared with those with a favorable outcome.23 A positive likelihood ratio is calculated as [sensitivity/(1–specificity)] and a negative likelihood ratio as [1–(sensitivity/specificity)].23 Therefore, the likelihood ratios are more informative than sensitivity and specificity alone, because the posttest probabilities are calculated from both sensitivity and specificity. In addition, likelihood ratios are less influenced by variation in prevalence of particular outcomes compared with predictive values. We sought to evaluate the positive and negative predictive value and likelihood ratios of discharge destination for an unfavorable outcome at 3 and 12 months in 2 strata defined by age (<65y and ≥65y) to remain consistent with age strata used in previous studies using age strata derived from administrative datasets.24 All multivariate analyses were performed using SAS version 9.0.a
RESULTS
Of the 624 patients recruited, 94 patients died during hospitalization and 530 patients were discharged from the hospital after ischemic stroke. Of these 530 patients, 242 were discharged home, 21 to a relative’s/friend’s home, 206 to a rehabilitation facility, and 61 to a nursing home. The mean time ±SD from recruitment to discharge was 12±11 days. The demographic and clinical characteristics of the patients according to discharge destination are provided in table 1. A higher proportion of patients aged ≥65 years and those with initial NIHSS ≥20 were seen in patients discharged to nursing homes. Table 1 also provides the rates of various categories of the MRS, and the proportions of patients with Barthel Index scores of 95 to 100 and Glasgow Outcome Scale scores of 0 to 1 in patients groups defined by discharge destination.
Table 1.
Demographic and Clinical Characteristics of Patients Discharged Alive After Ischemic Stroke According to a Discharge Destination
| Variables | Discharged to Home (n=242) n (%) |
Discharged to Relative’s/Friends’ Home (n=21) n (%) |
Discharged to Rehabilitation Facility (n=206) n (%) |
Discharged to Nursing Home (n=61) n (%) |
|---|---|---|---|---|
| Demographic characteristics | ||||
| Age strata (y) | ||||
| <65 | 116 (47.9) | 3 (14.3) | 75 (36.4) | 10 (16.4) |
| ≥65 | 126 (52.1) | 18 (85.7) | 131 (63.6) | 51 (83.6) |
| Men | 153 (63.2) | 8 (38.1) | 113 (54.8) | 29 (47.5) |
| Race/ethnicity | ||||
| Black | 76 (31.4) | 5 (23.8) | 53 (25.7) | 7 (11.5) |
| White | 146 (60.3) | 15 (71.4) | 136 (66.0) | 47 (77.1) |
| Hispanic | 12 (4.9) | 0 | 14 (6.8) | 5 (8.2) |
| Asian | 6 (2.5) | 1 (4.7) | 0 | 1 (1.6) |
| Other | 2 (0.83) | 0 | 3 (1.5) | 1 (1.6) |
| Clinical characteristics | ||||
| Baseline NIHSS strata | ||||
| 0–9 | 118 (48.7) | 8 (38.1) | 41 (19.9) | 6 (9.8) |
| 10–19 | 101 (41.7) | 9 (42.8) | 115 (55.8) | 27 (44.3) |
| ≥20 | 23 (9.5) | 4 (19.1) | 50 (24.3) | 28 (45.9) |
| Presumptive diagnosis | ||||
| Large vessel occlusion | 46 (19.0) | 3 (14.3) | 24 (11.6) | 3 (4.9) |
| Cardioembolic | 98 (40.5) | 10 (4.5) | 88 (42.7) | 27 (44.3) |
| Small vessel occlusion | 90 (37.2) | 8 (38.1) | 88 (42.7) | 30 (49.2) |
| Other | 8 (3.3) | 0 | 6 (2.9) | 1 (1.6) |
| Prior stroke at baseline | 35 (14.6) | 5 (25.0) | 20 (9.8) | 11 (18.0) |
| No preexisting disability at baseline | 42 (18.1) | 4 (21.0) | 29 (15.3) | 9 (16.8) |
| Diabetes mellitus | 46 (19.0) | 1 (4.7) | 40 (19.6) | 20 (33.3) |
| Hypertension | 148 (61.4) | 17 (80.9) | 138 (67.6) | 36 (61.0) |
| Myocardial infarction | 56 (24.0) | 4 (21.0) | 36 (18.5) | 12 (20.6) |
| Atrial fibrillation | 33 (13.8) | 5 (23.8) | 38 (18.5) | 13 (21.3) |
| Angina | 55 (23.2) | 3 (15.0) | 37 (18.8) | 13 (22.0) |
| Congestive heart failure | 31 (13.2) | 6 (31.5) | 29 (14.6) | 12 (21.4) |
| Valvular heart disease | 18 (7.6) | 4 (21.0) | 12 (6.1) | 4 (7.0) |
| Cigarette smoking | 105 (44.3) | 5 (23.8) | 60 (29.4) | 14 (23.3) |
| Hyperlipidemia | 57 (26.8) | 6 (40.0) | 45 (25.8) | 10 (21.3) |
| Outcome measures | ||||
| Barthel Index at 3mo | ||||
| <95 | 45 (19.8) | 9 (47.4) | 131 (65.8) | 51 (86.4) |
| ≥95 | 182 (80.2) | 10 (52.6) | 68 (34.2) | 8 (13.5) |
| Glasgow Outcome Scale Score at 3mo | ||||
| ≤1 | 175 (77.1) | 10 (52.6) | 42 (21.1) | 4 (6.7) |
| >1 | 52 (22.9) | 9 (47.3) | 157 (78.8) | 55 (93.2) |
| Outcome of patients at 3mo by the MRS | ||||
| 0 | 82 (33.8) | 5 (23.8) | 2 (0.9) | 0 |
| 1 | 93 (38.4) | 6 (28.5) | 21 (10.2) | 3 (4.9) |
| 2 | 30 (12.4) | 1 (4.7) | 28 (13.5) | 2 (3.3) |
| 3 | 14 (5.8) | 5 (23.8) | 55 (26.7) | 5 (8.2) |
| 4 | 8 (3.3) | 2 (9.5) | 70 (33.9) | 20 (32.8) |
| 5 | 7 (2.9) | 1 (4.7) | 15 (7.3) | 13 (21.3) |
| 6 | 8 (3.3) | 1 (4.7) | 15 (7.3) | 18 (29.5) |
| Outcome of patients at 12mo by the MRS | ||||
| 0 | 83 (36.5) | 5 (26.3) | 4 (2.0) | 1 (1.7) |
| 1 | 77 (33.9) | 4 (21.1) | 32 (16.1) | 3 (5.1) |
| 2 | 24 (10.5) | 2 (10.5) | 29 (14.5) | 3 (5.1) |
| 3 | 12 (5.3) | 2 (10.5) | 52 (26.1) | 6 (10.2) |
| 4 | 6 (2.6) | 1 (5.3) | 40 (20.1) | 6 (10.2) |
| 5 | 3 (1.3) | 2 (10.5) | 12 (6.0) | 14 (23.7) |
| 6 | 22 (9.7) | 3 (15.8) | 30 (15.1) | 26 (44.1) |
At 3-months poststroke, the rates of unfavorable outcome defined by the 3 MRS-based definitions were as follows: 2 to 6 (n=318), 3 to 6 (n=257), and 4 to 6 (n=178). There were 26 (4.2%) patients who were unavailable for the follow-up assessment at 12 months. Among the remaining 504 patients with 12-month follow-up, the rates of unfavorable outcome defined by the MRS definitions were as follows: 2 to 6 (n=295), 3 to 6 (n=237), and 4 to 6 (n=165). The positive and negative predictive value of discharge destination for unfavorable outcomes defined using the various MRS thresholds at 3 and 12 months is provided in table 2. The positive predictive value of discharge to nursing home or rehabilitation facility was the highest for unfavorable outcome defined by an MRS score of 2 to 6 (95% and 89%, respectively). The positive predictive value of rehabilitation facility/nursing home was also highest for unfavorable outcomes defined by an MRS score of 2 to 6 as compared with those defined by an MRS score of 3 to 6 (79%) and 4 to 6 (57%). The negative predictive value was the highest for unfavorable outcome defined by an MRS score of 4 to 6 (90%) compared with those defined by an MRS score of 3 to 6 (83%) and 2 to 6 (71%). A similar pattern was observed with unfavorable outcome, defined using various thresholds at 12 months.
Table 2.
Predictive Values and Likelihood Ratios for Categories of Discharge Destinations Alone or in Combination for (unfavorable) Outcome Defined by the MRS at 3 and 12 Months
| Discharge Destinations | Positive Predictive Value (%) | Negative Predictive Value (%) | Positive Likelihood Ratio (95% CI) | Negative Likelihood Ratio (95% CI) | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|
|
|
|
|||||||||
| MRS 2–6 | MRS 3–6 | MRS 4–6 | MRS 2–6 | MRS 3–6 | MRS 4–6 | MRS 2–6 | MRS 3–6 | MRS 4–6 | MRS 2–6 | MRS 3–6 | MRS 4–6 | |
| Unfavorable Outcome Defined by the MRS at 3mo | ||||||||||||
| Home | 28 | 15 | 10 | 13 | 24 | 46 | 0.3 (0.2–0.3) | 0.2 (0.1–0.3) | 0.21 (0.1–0.3) | 4.5 (3.3–6.1) | 3.4 (2.7–4.2) | 2.3 (1.9–2.7) |
| Relative’s/friend’s home | 48 | 43 | 19 | 39 | 51 | 66 | 0.6 (0.3–1.4) | 0.8 (0.3–1.9) | 0.47 (0.2–1.4) | 1.0 (0.9–1.1) | 1.0 (0.9–1.0) | 1.0 (0.9–1.0) |
| Rehabilitation facility | 89 | 75 | 49 | 58 | 69 | 76 | 5.3 (3.5–7.9) | 3.2 (2.4–4.2) | 1.87 (1.5–2.3) | 0.5 (0.4–0.5) | 0.4 (0.4–0.5) | 0.6 (0.5–0.7) |
| Nursing home | 95 | 92 | 84 | 45 | 57 | 73 | 13 (4.1–41) | 12 (4.8–29) | 10 (5.3–19) | 0.8 (0.7–0.8) | 0.8 (0.7–0.8) | 0.7 (0.7–0.8) |
| Rehabilitation facility/nursing home discharge | 90 | 79 | 57 | 71 | 83 | 90 | 6.2 (4.3–8.9) | 4.0 (3.1–5.1) | 2.6 (2.2–3.0) | 0.3 (0.2–0.3) | 0.2 (0.2–0.3) | 0.2 (0.2–0.3) |
| Unfavorable Outcome Defined by the MRS at 12mo | ||||||||||||
| Home | 23 | 18 | 19 | 18 | 30 | 52 | 0.3 (0.2–0.3) | 0.2 (0.2–0.3) | 0.3 (0.2–0.4) | 3.3 (2.5–4.2) | 2.6 (2.2–3.2) | 1.9 (1.6–2.2) |
| Relative’s/friend’s home | 53 | 42 | 32 | 41 | 53 | 67 | 0.8 (0.3–1.9) | 0.8 (0.3–2.0) | 0.9 (0.3–2.5) | 1.0 (0.9–1.0) | 1.0 (0.97–1.0) | 1.0 (0.9–1.0) |
| Rehabilitation facility | 82 | 67 | 41 | 57 | 66 | 73 | 3.2 (2.3–4.4) | 2.3 (1.8–2.9) | 1.4 (1.2–1.8) | 0.5 (0.5–0.6) | 0.5 (0.5–0.7) | 0.7 (0.6–0.91) |
| Nursing home | 93 | 88 | 78 | 46 | 58 | 73 | 9.7 (3.6–26) | 8.3 (3.8–18) | 7.2 (4.0–13) | 0.8 (0.8–0.8) | 0.8 (0.7–0.8) | 0.7 (0.6–0.8) |
| Rehabilitation facility/nursing home discharge | 84 | 72 | 50 | 69 | 79 | 85 | 3.8 (2.9–5.1) | 2.9 (2.4–3.5) | 2.0 (1.7–2.3) | 0.3 (0.3–0.4) | 0.3 (0.2–0.4) | 0.4 (0.2–0.5) |
The positive likelihood ratio was highest for nursing home discharges (13; 95% confidence interval [CI], 4.1– 41) followed by rehabilitation facility discharges for unfavorable outcome defined by an MRS score of 2 to 6 at 3-months poststroke (5.3; 95% CI, 3.5–7.9). A similar positive likelihood ratio was seen for unfavorable outcome defined by an MRS score of 3 to 6. The negative likelihood ratio was the highest for home discharge for unfavorable outcome defined by an MRS score of 2 to 6 (4.5; 95% CI, 3.3– 6.1). A similar pattern was observed with unfavorable outcomes defined using various thresholds at 12 months.
When the results were analyzed in 2 strata defined by age, the positive likelihood ratio for all discharge destinations for unfavorable outcome (MRS score 2–6) were similar between those aged less than 65 years compared with those aged ≥65 years (data not shown) (see supplemental tables 1 and 2, available online at the Archives website: www.archives-pmr.org).
DISCUSSION
Salient Findings
We found that discharge destination after acute care hospitalization had the highest positive predictive value and likelihood ratios for unfavorable outcome defined by an MRS score of 2 to 6 at 3-months poststroke compared with unfavorable outcome defined by other MRS thresholds. Discharge to nursing home provided the best positive predictive value for unfavorable outcome, although combining both discharge to nursing home or rehabilitation facility yielded similar results. Any home discharge (home or relative’s/friend’s home) provided the lowest negative predictive value for unfavorable outcome followed by home discharge when unfavorable outcome was defined by an MRS score of 2 to 6. The discharge destination has certain unique attributes that allow generalization and widespread use. Such attributes include: (1) availability of the data at the time of discharge and with no impact of attrition because of loss to follow-up; (2) comprehensible for patients and family to understand; (3) data are readily available through multitude of sources; and (4) ascertainment is not subject to interrater variability.
Our goal is not to demonstrate discharge destination as a more effective predictor of outcome after acute stroke. Functional measures, such as the MRS, the Glasgow Outcome Scale, and the Barthel Index, have been used as the primary outcome measures in acute stroke treatment trials. There have also been studies showing the usefulness of the NIHSS in categorizing stroke outcomes.25,26 On a comparative basis, an improvement to an NIHSS of 1 or less or by 11 or more points (prognosis adjusted) at 90 days turned out to be the most powerful outcome measure in a simulated study of 6000 acute stroke clinical trials.27 The resources required for such ascertainment may not always be possible for assessment in real world scenarios. Discharge destination appears to be an optimal surrogate measure for such scenarios.
The issues regarding the clinical and research implications of the findings need to be discussed. The effect of any programmatic change or introduction of new therapeutic interventions in such settings is best ascertained by rate of unfavorable outcome (defined by MRS) at 3-months poststroke in the patient population under study. However, there are numerous scenarios where data regarding 3-month ascertainment using the MRS are not available because of the lack of standardized follow-up within medical systems, infrastructure to collect clinical data, and/or patient compliance. Discharge disposition is readily available through medical records and ascertained consistently in administrative datasets.7–11 Our results allow extrapolation of outcomes defined by discharge destination into outcomes defined by an MRS score of 2 to 6 at 3 and 12 months. Other definitions of unfavorable outcome, such as an MRS score of 3 to 6 and 4 to 6, although used in previous studies,20–22 are less optimal for such extrapolation. Based on a positive predictive value of 95%, we can assume that 95 of 100 patients discharged to nursing home will have an MRS score of 2 to 6. But the negative predictive value is 45%, meaning that 55 of the 100 patients who were not discharged to a nursing home will also have an MRS score of 2 to 6. Therefore, if the researcher is interested in presence of unfavorable outcome in the study sample,28 the selection of patients with nursing home discharges may be the best option. However, if the researcher is interested in absence of unfavorable outcome in the study sample, then the goal may be best achieved by choosing discharge to home with a negative predictive value of 13% as the selection variable; 87% of persons who were discharged home are not going to have an MRS score of 2 to 6 at 3-months poststroke. Measures, such as the likelihood ratio, can assist in providing easily comprehensible prognostic information regarding chances of disability and death at the time of discharge based on destination. For example, the odds are 6 times higher in patients discharged to a nursing home (positive likelihood ratio=6) to have an MRS score of 2 to 6 at 3-months poststroke compared with those who were not discharged to nursing homes. Similarly, the odds are 4 times higher for patients discharged to home (negative likelihood ratio=4) to not have an MRS score of 2 to 6 at 3-months poststroke.
Study Limitations
There are certain issues that need to be understood prior to interpretation of our results. The predictive values are highly dependent on the pretest probability, which is 60% (MRS score of 2 to 6 was seen in 318 of 530 patients) in this sample of ischemic stroke patients. If the pretest probability was higher in patients with a diagnosis such as intracerebral or subarachnoid hemorrhage, the posttest probability would be higher. The data of derivation is based on a clinical trial, which is likely to consist of patients with more favorable characteristics and thus a lower rate of death and disability.29 The patients recruited in the NINDS rt-PA study were those who presented within 3 hours of symptom onset. In a population based study of 1590 patients with ischemic stroke,30 patients who presented earlier after symptom onset had more severe neurologic deficits (higher NIHSS) and lower rates of favorable outcome. Depending on the influence of various factors that influence the rate of death and disability in patients treated outside clinical trials, discharge destination to a nursing home or rehabilitation facility may have a higher or lower probability of predicting an MRS score of 2 to 6 in unselected groups of ischemic stroke patients. Discharge destination is not always based on functional status and magnitude of disability. Other factors such as patient-related factors, including social support and fiscal status, and institutional factors, such as inpatient rehabilitation capabilities, determine discharge destination.31 Discharge destination may underestimate the effect of posthospital improvement in functional status. Our analysis is derived from an aggregate sample of various ischemic stroke subtypes. The rate of death and disability varied substantially between patients with total anterior circulation infarcts and those with lacunar or partial anterior circulation infarcts in 1 study.32 Thus our results may be more reflective of an aggregate of stroke subtypes and may lose some of its validity in a patient population consisting of 1 stroke subtype.
CONCLUSIONS
Our results provide insight into the predictive value of discharge destination as a surrogate for defining favorable and unfavorable outcome at 3 and 12 months. These results are expected to allow better study design and interpretation for studies and clinical decisions that rely on discharge destination as an outcome.
Supplementary Material
Acknowledgments
Supported by the National Institutes of Health (grant no. RO-1-NS44976-01A2, 1U01NS062091-01A2, and 5K12-RR023247-05), ESP Pharma Inc., an American Heart Association Established Investigator Award (award no. 0840053N), and the Minnesota Medical Foundation, Minneapolis, MN.
List of Abbreviations
- CI
confidence interval
- MRS
Modified Rankin Scale
- NINDS
National Institute of Neurological Disorders and Stroke
- NIHSS
National Institutes of Health Stroke Scale score
- rt-PA
recombinant tissue plasminogen activator
Footnotes
SAS Institute, 100 SAS Campus Dr, Cary, NC 27513-2414.
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.
References
- 1.Mayer SA, Brun NC, Begtrup K, et al. Recombinant activated factor VII for acute intracerebral hemorrhage. N Engl J Med. 2005;352:777–85. doi: 10.1056/NEJMoa042991. [DOI] [PubMed] [Google Scholar]
- 2.Mayer SA, Brun NC, Begtrup K, et al. Efficacy and safety of recombinant activated factor VII for acute intracerebral hemorrhage. N Engl J Med. 2008;358:2127–37. doi: 10.1056/NEJMoa0707534. [DOI] [PubMed] [Google Scholar]
- 3.Shuaib A, Lees KR, Lyden P, et al. NXY-059 for the treatment of acute ischemic stroke. N Engl J Med. 2007;357:562–71. doi: 10.1056/NEJMoa070240. [DOI] [PubMed] [Google Scholar]
- 4.Wahlgren N, Ahmed N, Davalos A, et al. Thrombolysis with alteplase 3–4. 5 h after acute ischaemic stroke (SITS-ISTR): an observational study. Lancet. 2008;372:1303–9. doi: 10.1016/S0140-6736(08)61339-2. [DOI] [PubMed] [Google Scholar]
- 5.Wahlgren N, Ahmed N, Davalos A, et al. Thrombolysis with alteplase for acute ischaemic stroke in the Safe Implementation of Thrombolysis in Stroke-Monitoring Study (SITS-MOST): an observational study. Lancet. 2007;369:275–82. doi: 10.1016/S0140-6736(07)60149-4. [DOI] [PubMed] [Google Scholar]
- 6.Albers GW, Bates VE, Clark WM, Bell R, Verro P, Hamilton SA. Intravenous tissue-type plasminogen activator for treatment of acute stroke: the Standard Treatment with Alteplase to Reverse Stroke (STARS) study. JAMA. 2000;283:1145–50. doi: 10.1001/jama.283.9.1145. [DOI] [PubMed] [Google Scholar]
- 7.Qureshi AI, Suri MF, Nasar A, et al. Changes in cost and outcome among US patients with stroke hospitalized in 1990 to 1991 and those hospitalized in 2000 to 2001. Stroke. 2007;38:2180–4. doi: 10.1161/STROKEAHA.106.467506. [DOI] [PubMed] [Google Scholar]
- 8.Khatri R, Memon MZ, Zacharatos H, et al. Impact of percutaneous transluminal angioplasty for treatment of cerebral vasospasm on subarachnoid hemorrhage patient outcomes. Neurocrit Care. 2011;15:28–33. doi: 10.1007/s12028-010-9499-y. [DOI] [PubMed] [Google Scholar]
- 9.Qureshi AI, Suri MF, Nasar A, et al. Thrombolysis for ischemic stroke in the United States: data from National Hospital Discharge Survey 1999–2001. Neurosurgery. 2005;57:647–54. [PubMed] [Google Scholar]
- 10.Johnston SC. Effect of endovascular services and hospital volume on cerebral aneurysm treatment outcomes. Stroke. 2000;31:111–7. doi: 10.1161/01.str.31.1.111. [DOI] [PubMed] [Google Scholar]
- 11.Bardach NS, Zhao S, Gress DR, Lawton MT, Johnston SC. Association between subarachnoid hemorrhage outcomes and number of cases treated at California hospitals. Stroke. 2002;33:1851–6. doi: 10.1161/01.str.0000019126.43079.7b. [DOI] [PubMed] [Google Scholar]
- 12.The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med. 1995;333:1581–7. doi: 10.1056/NEJM199512143332401. [DOI] [PubMed] [Google Scholar]
- 13.Kwiatkowski TG, Libman RB, Frankel M, et al. Effects of tissue plasminogen activator for acute ischemic stroke at one year. National Institute of Neurological Disorders and Stroke Recombinant Tissue Plasminogen Activator Stroke Study Group. N Engl J Med. 1999;340:1781–7. doi: 10.1056/NEJM199906103402302. [DOI] [PubMed] [Google Scholar]
- 14.Patel SC, Levine SR, Tilley BC, et al. Lack of clinical significance of early ischemic changes on computed tomography in acute stroke. JAMA. 2001;286:2830–8. doi: 10.1001/jama.286.22.2830. [DOI] [PubMed] [Google Scholar]
- 15.The National Institute of Neurological Disorders and Stroke (NINDS) rt-PA Stroke Study Group. A systems approach to immediate evaluation and management of hyperacute stroke. Experience at eight centers and implications for community practice and patient care. Stroke. 1997;28:1530–40. doi: 10.1161/01.str.28.8.1530. [DOI] [PubMed] [Google Scholar]
- 16.van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke. 1988;19:604–7. doi: 10.1161/01.str.19.5.604. [DOI] [PubMed] [Google Scholar]
- 17.Banks JL, Marotta CA. Outcomes validity and reliability of the modified Rankin scale: implications for stroke clinical trials: a literature review and synthesis. Stroke. 2007;38:1091–6. doi: 10.1161/01.STR.0000258355.23810.c6. [DOI] [PubMed] [Google Scholar]
- 18.Sulter G, Steen C, De Keyser J. Use of the Barthel index and modified Rankin scale in acute stroke trials. Stroke. 1999;30:1538–41. doi: 10.1161/01.str.30.8.1538. [DOI] [PubMed] [Google Scholar]
- 19.Bluhmki E, Chamorro A, Davalos A, et al. Stroke treatment with alteplase given 3.0–4. 5 h after onset of acute ischaemic stroke (ECASS III): additional outcomes and subgroup analysis of a randomised controlled trial. Lancet Neurol. 2009;8:1095–102. doi: 10.1016/S1474-4422(09)70264-9. [DOI] [PubMed] [Google Scholar]
- 20.Furlan A, Higashida R, Wechsler L, et al. Intra-arterial prourokinase for acute ischemic stroke. The PROACT II study: a randomized controlled trial. Prolyse in Acute Cerebral Thromboembolism. JAMA. 1999;282:2003–11. doi: 10.1001/jama.282.21.2003. [DOI] [PubMed] [Google Scholar]
- 21.Smith WS, Sung G, Saver J, et al. Mechanical thrombectomy for acute ischemic stroke: final results of the Multi MERCI trial. Stroke. 2008;39:1205–12. doi: 10.1161/STROKEAHA.107.497115. [DOI] [PubMed] [Google Scholar]
- 22.Vahedi K, Hofmeijer J, Juettler E, et al. Early decompressive surgery in malignant infarction of the middle cerebral artery: a pooled analysis of three randomised controlled trials. Lancet Neurol. 2007;6:215–22. doi: 10.1016/S1474-4422(07)70036-4. [DOI] [PubMed] [Google Scholar]
- 23.Sonis J. How to use and interpret interval likelihood ratios. Fam Med. 1999;31:432–7. [PubMed] [Google Scholar]
- 24.Brinjikji W, Rabinstein AA, Kallmes DF, Cloft HJ. Patient outcomes with endovascular embolectomy therapy for acute ischemic stroke: a study of the national inpatient sample: 2006 to 2008. Stroke. 2011;42:1648–52. doi: 10.1161/STROKEAHA.110.607952. [DOI] [PubMed] [Google Scholar]
- 25.Adams HP, Jr, Davis PH, Leira EC, et al. Baseline NIH Stroke Scale score strongly predicts outcome after stroke: a report of the Trial of Org 10172 in Acute Stroke Treatment (TOAST) Neurology. 1999;53:126–31. doi: 10.1212/wnl.53.1.126. [DOI] [PubMed] [Google Scholar]
- 26.Broderick JP, Lu M, Kothari R, et al. Finding the most powerful measures of the effectiveness of tissue plasminogen activator in the NINDS tPA stroke trial. Stroke. 2000;31:2335–41. doi: 10.1161/01.str.31.10.2335. [DOI] [PubMed] [Google Scholar]
- 27.Young FB, Weir CJ, Lees KR. Comparison of the National Institutes of Health Stroke Scale with disability outcome measures in acute stroke trials. Stroke. 2005;36:2187–92. doi: 10.1161/01.STR.0000181089.41324.70. [DOI] [PubMed] [Google Scholar]
- 28.Department of Clinical Epidemiology and Biostatistics. McMaster University Health Sciences Centre. How to read clinical journals: II. To learn about a diagnostic test. Can Med Assoc J. 1981;124:703–10. [PMC free article] [PubMed] [Google Scholar]
- 29.Qureshi AI, Hutson AD, Harbaugh RE, Stieg PE, Hopkins LN. Methods and design considerations for randomized clinical trials evaluating surgical or endovascular treatments for cerebrovascular diseases. Neurosurgery. 2004;54:248–64. doi: 10.1227/01.neu.0000103446.26057.78. discussion 264–7. [DOI] [PubMed] [Google Scholar]
- 30.Qureshi AI, Kirmani JF, Sayed MA, et al. Buffalo metropolitan area and Erie County stroke study: rationale, design, and methods. Neuroepidemiology. 2004;23:289–98. doi: 10.1159/000080095. [DOI] [PubMed] [Google Scholar]
- 31.Rundek T, Mast H, Hartmann A, et al. Predictors of resource use after acute hospitalization: the Northern Manhattan Stroke Study. Neurology. 2000;55:1180–7. doi: 10.1212/wnl.55.8.1180. [DOI] [PubMed] [Google Scholar]
- 32.Bamford J, Sandercock P, Dennis M, Burn J, Warlow C. Classification and natural history of clinically identifiable subtypes of cerebral infarction. Lancet. 1991;337:1521–6. doi: 10.1016/0140-6736(91)93206-o. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.