Abstract
Measurement of quality of stroke care has become increasingly important, but data come mostly from programs in hospitals that choose to participate in certification programs, which may not be representative of the care provided in nonparticipating hospitals. The authors sought to determine differences in quality of care metric concordance for acute ischemic stroke among hospitals designated as a primary stroke center, comprehensive stroke center, and non-stroke center in a population-based epidemiologic study. Significant differences were found in both patient demographics and in concordance with guideline-based quality metrics. These differences may help inform quality improvement efforts across hospitals involved in certification as well as those that are not.
Keywords: stroke, ischemic stroke, certification, guidelines, guideline concordance
Stroke center (SC) certification programs have been associated with lower mortality1 and improved functional outcomes,2-4 which may be related to both hospital- and patient-level factors.5 It has been suggested that this improvement may be related in part to greater concordance with stroke quality metrics and guidelines, but data are limited,6 especially because there is a dearth of data from hospitals that do not participate in SC certification programs.
Existing quality of care data come largely from “before and after” evaluations of hospitals that choose to seek certification status, which may introduce volunteer bias and limit representativeness.7,8 National quality data sets and public “report cards” tend to be derived from administrative databases that indirectly measure guideline concordance and often lack detailed individual patient data,6,9 including stroke severity3—an important predictor of outcomes.10 Overall, there are very limited data regarding the quality of care provided to patients who are treated at hospitals that do not choose to become certified SCs. Despite this, it is critically important to consider hospitals both with and without SC certification when identifying areas for improvement in stroke care. Even in the United States, most hospitals do not have resources to become a primary SC (PSC).7 As of 2014, approximately one third of the US population did not live within a 60-minute drive of a certified SC,11 with significant disparities in this access, especially for rural individuals and residents of the southeastern “Stroke Belt.”12 Access to quality stroke care is an issue of international import for regions both with and without access to advanced stroke care.13
The Greater Cincinnati/Northern Kentucky Stroke Study (GCNKSS) is a population-based epidemiologic study that includes patient-level demographic, risk factor, and clinical details. It reflects care provided at all regional hospitals, both with and without SC certification, in a biracial (black and white) population of 1.3 million. We sought to use data from this study to compare ischemic stroke (IS) guideline concordance between hospitals in the same community designated as non-stroke center (nSC), PSC, and comprehensive stroke center (CSC).
Methods
The full methods of the GCNKSS, a population-based epidemiologic study, have been described previously.14 Briefly, records were reviewed from all acute stroke visits to all regional acute care hospitals and emergency departments (EDs), as defined by discharge International Classification of Diseases, Ninth Revision, diagnoses. Charts were abstracted by stroke study nurses to include all relevant clinical data, including demographics, presenting symptoms, examination findings, risk factors, medication use, laboratory and imaging results, and hospital providing treatment. A stroke-trained study physician reviewed these chart abstractions to ascertain whether a stroke or transient ischemic attack occurred, and to identify stroke subtype (IS, intracranial hemorrhage, subarachnoid hemorrhage, transient ischemic attack) and presumed mechanism. National Institutes of Health Stroke Scale (NIHSS) scores were determined retrospectively using previously described methods in which NIHSS data not explicitly recorded (eg, whether the patient was able to answer NIHSS-specific orientation questions) were extrapolated based on documented aspects of the physical exam (eg, documentation of “oriented ×3”).15 The GCNKSS was approved by the institutional review boards of all participating institutions.
Hospital and Patient Characteristics
The current investigation is a retrospective analysis of adult (age ≥20 years) patients among residents of the Greater Cincinnati/Northern Kentucky region who had acute IS in 2010, excluding those that occurred during hospitalization. Patient-level data were abstracted and cross-referenced to harmonized stroke performance measures that were considered best practice during the study period. All regional hospitals, excluding the local Veterans Health Administration and Children’s Hospital, were grouped according to their Joint Commission (TJC) SC certification status during the study period. Ten hospitals were classified as nSC and 2 as PSC; 1 additional PSC achieved CSC certification when that designation was implemented in 2013. As of 2017, this hospital remains the only CSC in the region, and it has remained the only academic medical center and the preferred transfer destination for complex stroke care in the region. The study team chose to consider this hospital separately (“future CSC”) to evaluate to what extent its care patterns may have differed from hospitals that maintained PSC status but did not seek CSC certification. One hospital that obtained PSC certification status in October 2010 was grouped with the nSC, as this was its status for the majority of the study period. No regional hospitals were stroke certified by bodies other than TJC.
Patient demographic data and guideline-relevant risk factors were gathered, including age, race, sex, diagnosis of atrial fibrillation or flutter, and current smoking status. Time to patient arrival to ED after last known to be well was ascertained based on the time of witnessed onset or, if onset time was unknown, the time between last known to be well and ED arrival.16 Hyperlipidemia was defined as patients with low-density lipoprotein cholesterol (LDL) ≥2.6 mmol/L (100 mg/dL) or on a lipid-lowering therapy prior to arrival.
Guideline Concordance and Outcomes
Guideline concordance was evaluated according to the harmonized consensus stroke performance measures developed in 2008 by representatives from the Centers for Disease Control and Prevention, the American Heart Association’s Get With the Guidelines-Stroke, TJC, and the National Quality Forum (NQF).17 Among these harmonized measures, the only discrepancies for IS evaluation were that NQF did not include dysphagia screening, and NQF evaluated smoking cessation globally, rather than as a stroke-specific measure17; the study team chose to include both of these in the analysis. Lipid-lowering therapy with statins or non-statin agents was considered appropriate, consistent with guidelines during the study period. The team did not analyze compliance with stroke education or timing of antithrombotic therapy, as neither was readily ascertainable from the GCNKSS data set.
The study team analyzed concordance with the following guidelines: intravenous (IV) recombinant tissue plasminogen activator (rtPA) within 3 hours among those arriving to the ED within 2 hours since last known to be well; venous thromboembolism (VTE) prophylaxis (chemoprophylaxis or mechanical) among nonambulatory patients; antithrombotic therapy prior to discharge; anticoagulation therapy at discharge for those with diagnosis of atrial fibrillation or flutter; lipid-lowering therapy among those with hyperlipidemia, LDL unmeasured during hospitalization, or those admitted on lipid-lowering therapy; dysphagia screening during hospitalization; smoking cessation counseling among smokers; rehabilitation assessment during hospitalization. For patients transferred between hospitals, arrival within 2 hours, initial NIHSS score, and time to IV rtPA treatment were attributed to the presenting hospital. The remainder of demographic and guideline concordance data were attributed to the final hospital.
When evaluating care for guideline concordance, patients were excluded from most analyses if they were enrolled in a clinical trial, or if they were made comfort care, hospice care, left against medical advice, or died during their acute stroke hospitalization. VTE prophylaxis measurements excluded those who were discharged, receiving comfort care, left against medical advice, or deceased prior to hospital day 2; if patients did not meet these criteria by day 2, VTE prophylaxis was considered a component of guideline-concordant care. Documentation of dysphagia evaluation was considered guideline concordant for all patients who were not NPO (nothing by mouth) for their entire hospital stay, regardless of the foregoing. Results were assessed based on whether or not the intervention was provided; documentation regarding appropriateness of nontreatment was not available. Outcome data were analyzed with regard to the rates of death at discharge and at 7 and 30 days.
Data Analysis
Differences in patient characteristics and hospital guideline concordance were evaluated using χ2 or Fisher exact test as appropriate. Differences in age and NIHSS were evaluated using the Kruskal-Wallis test. Unadjusted outcome analyses were performed using the χ2 test. Logistic regression was used to adjust for age and presenting NIHSS.
Results
A total of 1981 acute IS cases (median age 71 years, IQR 58-82; 55% women; 22% black) were identified during the study period. Of these patients, 1417 (71.5%) presented to an nSC, 323 (16.3%) to a PSC, and 241 (12.2%) to the future CSC. Patient transfers included 32 from nSC to PSC, 44 nSC to CSC, 7 PSC to CSC, and 13 nSC to another nSC.
Table 1 shows demographic and patient characteristics by presentation or destination hospital, as appropriate. Significant differences in average NIHSS by presenting hospital were greater when calculating by destination hospital; patients with more severe strokes were treated at a SC. A significant difference in the percentage of patients presenting to hospital within 2 hours of last known to be well was present only when considering destination hospital, indicating transfers of early-arriving patients to a SC.
Table 1.
Demographics and Patient Characteristics by Hospital Type, GCNK Region, 2010.
| Future CSC | PSC | nSC | P* | |
|---|---|---|---|---|
| Number of hospitals | 1 | 2 | 10 | — |
| Patient characteristics by presenting hospital (before transfers) | ||||
| Number of patients | 241 | 323 | 1417 | — |
| Last well to ED 0-2 hours | 57 (24%) | 67 (21%) | 273 (19%) | 0.27 |
| Initial NIHSS score, median (IQR) [mean (SD)] | 3 (1, 9)A [7.0 (8.2)] | 3 (1, 7)A [6.1 (7.9)] | 3 (1, 6) [5.3 (7.1)] | <.01 |
| Demographic and patient characteristics by final hospital (after transfers) | ||||
| Number of patients | 292 | 348 | 1341 | — |
| Median age (IQR) | 62 (54, 73) | 69 (58, 80) | 73 (61, 83) | <.01 |
| Female sex | 138 (47%) | 192 (55%)A | 759 (57%)A | .01 |
| Black race | 130 (45%) | 9 (3%) | 284 (21%) | <.01 |
| Initial NIHSS score, median (IQR) [mean (SD)] | 4 (1, 13) [8.1 (9.0)] | 3 (1, 7) [5.9 (7.9)] | 3 (1, 6) [5.0 (6.8)] | <.01 |
| Atrial fibrillation or flutter | 51 (17%) | 84 (24%)A | 336 (25%)A | .02 |
| Hyperlipidemia | 205 (70%)AB | 263 (76%)A | 917 (68%)B | .03 |
| Current smoker | 120 (41%) | 91 (26%)A | 338 (25%)A | <.01 |
Abbreviations: CSC, comprehensive stroke center; ED, emergency department; GCNK, Greater Cincinnati/Northern Kentucky; IQR, interquartile range; NIHSS, National Institutes of Health stroke scale; nSC, non-stroke center; PSC, primary stroke center.
Note: Bold indicates statistically a significant difference between the bolded item and at least one other item in its row.
P value reflects between-group significance.
Items within a row that share a common superscript do not differ at P < .05.
Hospitals with SC certification had greater guideline concordance for time to rtPA, dysphagia screening, smoking cessation education, use of lipid-lowering therapy, and VTE prophylaxis, but there were no significant differences between SC and nSC concordance with antithrombotic therapy, anticoagulation for atrial fibrillation, or rehabilitation assessment (Table 2).
Table 2.
Guideline-Concordant Care by Hospital Type, GCNK Region, 2010.
| Future CSC | PSC | nSC | P* | |
|---|---|---|---|---|
| Number of hospitals | 1 | 2 | 10 | — |
| Quality measures by presenting hospital (before transfers) | ||||
| 3 hours to rtPA among onset to ED 0-2 hours | 21 (37%)A | 27 (40%)A | 47 (17%) | <.01 |
| Quality measures by final hospital (after transfers)† | ||||
| Dysphagia screen | 244 (84%)A | 274 (79%)A | 742 (55%) | <.01 |
| VTE prophylaxis or ambulatory | 244 (99%) | 300 (96%) | 1108 (92%) | <.01 |
| Antithrombotic therapy‡ | 236 (97%) | 295 (95%) | 1129 (95%) | .40 |
| Anticoagulation for atrial fibrillation or flutter‡ | 19 (58%) | 37 (55%) | 153 (59%) | .88 |
| Lipid-lowering therapy | 168 (84%)A | 205 (79%)AB | 779 (74%)B | <.01 |
| Smoking cessation counseling | 103 (94%)A | 80 (94%)A | 240 (77%) | <.01 |
| Assessed for rehabilitation | 209 (86%) | 265 (85%) | 1002 (84%) | .79 |
Abbreviations: CSC, comprehensive stroke center; ED, emergency department; GCNK, Greater Cincinnati/Northern Kentucky; Bold indicates statistically a significant difference between the bolded item and at least one other item in its row. nSC, non-stroke center; PSC, primary stroke center; rtPA, recombinant tissue plasminogen activator; VTE, venous thromboembolism.
P value reflects between-group significance.
Among patients eligible for each measure.
At discharge.
Percentages within a row that share a common superscript do not differ at P < .05.
Note: Bold indicates statistically a significant difference between the bolded item and at least one other item in its row.
When evaluating patients by presenting hospital, there were no statistically significant differences in death at discharge or 7 or 30 days. When evaluating by destination hospital, the CSC had a higher rate of death at discharge (CSC 7% vs PSC 3% and nSC 4%, P = .004) and at 7 days (CSC 9% vs PSC 4% and nSC 5%, P = .046); neither difference persisted after adjustment for age and initial NIHSS. There were no statistically significant differences in death at 30 days, before or after adjustment.
Discussion
This study found SC certification to be associated with increased adherence to guideline-based quality of care metrics. These results add population-based evidence to the previous limited data comparing guideline concordance between SC and nSC18 and may help explain observed differences in outcomes between hospitals of different certification types.1,2
These results can be used to identify areas of stroke care that have more room for improvement, leading to targeted interventions and improvements in care. For example, the data suggest that it may be more effective for hospitals of all types to target improvements in dysphagia screening, especially considering the high attributable mortality of pneumonia among stroke patients.19 Similarly, nSC might be targeted for improvements in smoking cessation counseling, whereas all hospitals might be targeted for improvements in the prescription of lipid-lowering therapy.
Interestingly, the future CSC demonstrated greater guideline concordance with VTE prophylaxis than PSCs. As both groups were working under the same certification metrics during the study period, this suggests that internal factors distinct from certification standards drove this difference.
In addition to differences in guideline concordance, significant differences also were found in all demographics and all measured patient characteristics between hospital types. Although these differences should not affect the subsequent guideline concordance of care provided, they highlight the importance of targeting interventions appropriate to populations and evaluating stroke outcomes in the context of appropriate risk adjustment. Given the relationship between stroke severity and stroke mortality,10 the present study data demonstrate the importance of risk adjustment by stroke severity when measuring outcomes.
After adjustment for age and NIHSS score, no significant differences were found in mortality outcome measures. This is not unexpected, given the relatively low rate of early mortality after stroke, and the mortality risk associated with factors not directly related to in-hospital post-stroke care, such as comorbidity burden and stroke severity.20
The present study has the benefit of detailed patient-level data, including comorbidities that are often absent from administrative data. The study team was not able to ascertain whether care or counseling were provided but not documented, a limitation of all chart-based studies. However, previous studies have suggested that guideline-based quality-of-care improvements relate to changes in care, rather than changes in documentation.21 The team was limited in the ability to further separate out the effects of unmeasured confounders between hospitals, such as practice location. Importantly, the team was unable to establish in which circumstances it was appropriate to withhold guideline-recommended care. This is likely the reason that anticoagulation use was relatively low in this study; it may have influenced other metrics as well.
In summary, this is the first population-based study not based on administrative data comparing guideline concordance between SCs and nSCs. Differences in guideline concordance, as well as those in patient characteristics, identify achievable areas of improvement for all hospital types. Targeted improvements in quality of care can affect hundreds of thousands of patients each year, including significant proportions of the population not treated at certified SCs. Programs and policies that aim to improve stroke care need to consider targeting interventions in hospitals of all types, including hospitals involved in SC certification and those that are not.
Acknowledgments
Declaration of Conflicting Interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr Jasne received modest financial support from Medtronic. Dr Flaherty received financial support from Janssen (significant) and CSL Behring (modest) and is a member of Portola’s speakers’ bureau. Dr de los Rios la Rosa is a member of Boehringer Ingelheim speakers’ bureau. Dr Kleindorfer received financial support from Genentech (modest) for speaker’s bureau fees. Dr Mackey is funded by IUH-VFR-365, IUH/IUSM Strategic Research Initiative, Indiana University CTSI KL2 (NIH, UL1TR001108), and CTSI PDT (ICTSI NIH/NCRR RR025761); he also is a National Institutes of Health Loan Repayment Program recipient. The rest of the authors declare that they have no conflicts of interest.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by National Institutes of Health National Institute of Neurological Disorders and Stroke Grant R01NS30678 and Stroke T-32 Cerebrovascular Fellowship Training Program for Cerebrovascular Disease T32NS047996.
References
- 1.Bekelis K, Marth NJ, Wong K, Zhou W, Birkmeyer JD, Skinner J. Primary stroke center hospitalization for elderly patients with stroke: implications for case fatality and travel times. JAMA Intern Med. 2016;176:1361–1368. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Song S, Fonarow GC, Olson DM, et al. Association of get with the guidelines-stroke program participation and clinical outcomes for Medicare beneficiaries with ischemic stroke. Stroke. 2016;47:1294–1302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Xian Y, Holloway RG, Chan PS, et al. Association between stroke center hospitalization for acute ischemic stroke and mortality. JAMA. 2011;305:373–380. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Meretoja A, Roine RO, Kaste M, et al. Effectiveness of primary and comprehensive stroke centers. Stroke. 2010; 41:1102–1107. [DOI] [PubMed] [Google Scholar]
- 5.Reeves MJ, Gargano J, Maier KS, et al. Patient-level and hospital-level determinants of the quality of acute stroke care. Stroke. 2010;41:2924–2931. [DOI] [PubMed] [Google Scholar]
- 6.Parker C, Schwamm LH, Fonarow GC, Smith EE, Reeves MJ. Stroke quality metrics. Stroke. 2012;43:155–162. [DOI] [PubMed] [Google Scholar]
- 7.Gorelick PB. Primary and comprehensive stroke centers: history, value and certification criteria. J Stroke. 2013;15:78–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Schwamm LH, Fonarow GC, Reeves MJ, et al. Get with the guidelines-stroke is associated with sustained improvement in care for patients hospitalized with acute stroke or transient ischemic attack. Circulation. 2009;119:107–115. [DOI] [PubMed] [Google Scholar]
- 9.Kelly A, Thompson JP, Tuttle D, Benesch C, Holloway RG. Public reporting of quality data for stroke: is it measuring quality? Stroke J Cereb Circ. 2008;39:3367–3371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Fonarow GC, Pan W, Saver JL, et al. Comparison of 30-day mortality models for profiling hospital performance in acute ischemic stroke with vs without adjustment for stroke severity. JAMA. 2012;308:257–264. [DOI] [PubMed] [Google Scholar]
- 11.Adeoye O, Albright KC, Carr BG, et al. Geographic access to acute stroke care in the United States. Stroke. 2014;45:3019–3024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Mullen MT, Wiebe DJ, Bowman A, et al. Disparities in accessibility of certified primary stroke centers. Stroke. 2014;45:3381–3388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lindsay P, Furie KL, Davis SM, Donnan GA, Norrving B. World Stroke Organization global stroke services guidelines and action plan. Int J Stroke. 2014;9(suppl A100):4–13. [DOI] [PubMed] [Google Scholar]
- 14.Broderick J, Brott T, Kothari R, et al. The Greater Cincinnati/Northern Kentucky stroke study. Stroke. 1998; 29:415–421. [DOI] [PubMed] [Google Scholar]
- 15.Williams LS, Yilmaz EY, Lopez-Yunez AM. Retrospective assessment of initial stroke severity with the NIH Stroke Scale. Stroke. 2000;31:858–862. [DOI] [PubMed] [Google Scholar]
- 16.Madsen TE, Sucharew H, Katz B, et al. Gender and time to arrival among ischemic stroke patients in the Greater Cincinnati/Northern Kentucky stroke study. J Stroke Cerebrovasc Dis. 2016;25:504–510. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Reeves MJ, Parker C, Fonarow GC, Smith EE, Schwamm LH. Development of stroke performance measures. Stroke. 2010;41:1573–1578. [DOI] [PubMed] [Google Scholar]
- 18.Johnson AM, Goldstein LB, Bennett P, O’Brien EC, Rosamond WD; Investigators of the Registry of the North Carolina Stroke Center Collaborative. Compliance with acute stroke care quality measures in hospitals with and without primary stroke center certification: the North Carolina stroke care collaborative. J Am Heart Assoc. 2014;3(2):e000423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Armstrong JR, Mosher BD. Aspiration pneumonia after stroke. Neurohospitalist. 2011;1:85–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Appelros P, Nydevik I, Viitanen M. Poor outcome after first-ever stroke: predictors for death, dependency, and recurrent stroke within the first year. Stroke. 2003;34:122–126. [DOI] [PubMed] [Google Scholar]
- 21.Reeves MJ, Grau-Sepulveda MV, Fonarow GC, Olson DM, Smith EE, Schwamm LH. Are quality improvements in the get with the guidelines: stroke program related to better care or better data documentation? Circ Cardiovasc Qual Outcomes. 2011;4:503–511. [DOI] [PubMed] [Google Scholar]