Stroke is the fourth leading cause of death in the United States.[1] Though actual stroke‐related death has declined nationally by 19.4%, stroke‐related morbidity is still a significant burden.[2, 3] Hospital certification programs have been developed to improve the quality of stroke care on state and national levels. The Brain Attack Coalition (BAC) proposed 2 levels of stroke hospitals: primary stroke centers (PSCs) and comprehensive stroke centers (CSCs).[3, 4] Although most stroke patients can be cared for at PSCs, CSCs are able to care for more complex stroke patients.[4, 5] Using BAC recommendations the Joint Commission (JC) and American Heart Association/American Stroke Association created a certification program for PSCs. Eight evidence‐based performance measures are currently required for JC PSC certification.[6]

At the state level, New Jersey and Florida began designating PSCs and CSCs.[7, 8] New Jersey PSC and CSC designation criteria incorporate the elements of JC PSC certification, despite preceding them by several years.[6, 9] New Jersey CSC certification consists of more comprehensive requirements (Table 1). All New Jersey‐designated stroke centers submit data in the New Jersey Acute Stroke Registry (NJASR),[7] which closely matches the Centers for Disease Control and Prevention's Paul Coverdell National Acute Stroke Registry and includes the JC‐required core measures.

There is a paucity of data comparing state‐designated CSCs and PSCs, largely because few states have state designation programs. Although a recent observational study from Finland showed better outcomes in patients treated at CSCs, measures in that study were limited to mortality and institutionalization at 1 year.[10] In this study, we examined adherence of all New Jersey state‐designated stroke centers to the JC PSC measures and compared CSCs to PSCs in this regard. We posited that better compliance with these evidence‐based measures might translate into better quality of stroke care in the state and may lend support to future, larger studies that may be conducted because of the recent certification of CSCs by the JC.

METHODS

Components of the NJASR, key components of PSCs and CSCs in the BAC report, and the 8 JC's core stroke measures for PSC certification were assessed.[3, 4, 6]

First responders in New Jersey are required to bring suspected stroke patients to the nearest stroke‐designated hospital, regardless of whether it is a PSC or CSC, unless the patient is too medically unstable and needs to be taken to the nearest hospital. From there, decisions can be made to transfer a patient to a higher level hospital (CSC) depending on the complexity of the patient's condition.

New Jersey state‐designated PSCs and CSCs are required to abstract patient‐level data, evaluate outcomes, and initiate quality improvement activities on all patients evaluated for ischemic stroke, hemorrhagic stroke, transient ischemic attack (TIA), and those who undergo acute interventional therapy. Data are submitted quarterly to the New Jersey Department of Health and Senior Services (NJDHSS).[7] Hospital data are imported into the Acute Stroke Registry Database. The NJASR statewide dataset used for this analysis included all stroke admissions for the calendar years 2010 and 2011 and contains patient demographic information, health history, clinical investigations performed, treatments provided, and outcome measures that allow for risk‐adjusted assessment of outcomes.

The JC core stroke measures (STK‐1 thru 10 [except STK‐7 and 9]: venous thromboembolism [VTE] prophylaxis, discharged on antithrombotic therapy, anticoagulation therapy for atrial fibrillation/flutter, thrombolytic therapy, antithrombotic therapy by the end of hospital day 2, discharged on statin medication, stroke education, and assessment for rehabilitation) (except STK‐7 [dysphagia screening] and STK‐9 [smoking cessation/advice counseling]) apply only to acute ischemic and/or hemorrhagic stroke patients.

In our analysis, transferred patients and patients with a diagnosis of TIA, stroke not otherwise specified, and nonstroke‐related diagnoses were excluded. Hospital identity was kept anonymous through assignment of random numeric codes by the NJDHSS. Hospitals were categorized as CSC or PSC based on NJDHSS designation. Stroke severity on admission was assessed by categorizing National Institutes of Health Stroke Scale (NIHSS) scores into: no stroke when NIHSS=0, mild stroke when NIHSS=14, and moderate to severe stroke when NIHSS>5. Median door‐to‐thrombolytic drug times were assessed for both patients who arrived to the hospital within 2 hours of stroke symptom onset and received thrombolytic therapy within 3 hours, as well as patients who arrived within 3.5 hours (210 minutes) and received treatment within 4.5 hours (270 minutes).

RESULTS

There were 36,892 acute stroke cases treated at the 53 New Jersey PSCs and 12 CSCs in the calendar year 2010 and 2011 (Table 2). Sixty percent were treated at PSCs and 40% at CSCs. There were significant differences in the distribution of patients' characteristics (race, age, and gender) between the 2 hospital levels. At both PSCs and CSCs, the majority of patients were white, distantly followed by blacks. Patients at PSCs were statistically significantly older than CSCs. The most prevalent comorbid conditions in both PSCs and CSCs were hypertension, diabetes mellitus, and dyslipidemia. Based on our categorization, we found that 45% of patients admitted to CSCs had moderate‐to‐severe stroke (NIHSS>5). The median door‐to‐thrombolytic drug times were significantly shorter at CSCs than PSCs for both the 3‐hour (65 vs 74 minutes, P<0.0001) and 4.5 hour (65 vs 76 minutes, P<0.0001) IV tPA time windows.

The incidences of stroke diagnosis types are also detailed in Table 2. Seventy percent of patients at CSCs had either an ischemic or hemorrhagic stroke diagnosis versus 57.1% of patients admitted at PSCs. Hemorrhagic stroke patients were twice as likely to be admitted at CSCs compared to PSCs.

After excluding 13,964 patients with a diagnosis of TIA, stroke not otherwise specified, and those with nonstroke‐related diagnosis, the likelihood of stroke patients' receiving the JC's performance measure services at either of these hospital levels was assessed (Table 3). In general, the adjusted odds ratio estimates of patients receiving a JC core performance measure at PSCs were lower than CSCs, indicating better compliance with the measures at CSCs. For example, 19.5% of eligible patients received thrombolytic therapy at CSCs compared to 9.6% at PSCs. CSCs also were more likely to provide VTE prophylaxis, anticoagulation for atrial fibrillation, and assessment for rehabilitation. Stroke education and antithrombotic therapy by the end of hospital day 2 were more likely to be provided at PSCs, but the results were not statistically significant.

DISCUSSION

In New Jersey, CSCs were more likely to adhere better to JC core performance measures than PSCs. Median door‐to‐thrombolytic drug times were also significantly lower at CSCs. Such differences may be due to several factors including the fact that CSCs have generally been state designated for a longer period of time than PSCs. CSCs are likely to have higher volumes of stroke admissions, are more likely to be JC certified, provide more staff education, and have more staff and resources. The New Jersey stroke designation program began in 2006, and 11 of the 12 CSCs were designated by the end of 2007. However, the PSC designation process has been more gradual, with several of them designated in 2010 and 2011 as the data for this study were being collected.

The New York State Stroke Center Designation Project prospectively showed that stroke center designation improved the quality of acute stroke patient care and administration of thrombolytic therapy; however, differing levels of hospital designation were not present in New York at that time.[11] Participation in a data measurement program such as Get With The Guidelines has also been examined. It is evident that the amount of time in a program is predictive of process measure compliance.[12] JC certification as a PSC is also associated with increased thrombolytic rates for acute stroke over time.[13] New Jersey does not require that stroke‐designated hospitals have JC stroke certification. Although 11 New Jersey CSCs have been certified as JC PSCs since 2009, only 21 of the 53 state‐designated PSCs are JC certified. It may be that the highest performing sites pursue state CSC designation and JC PSC certification/recertification repeatedly. CSCs in New Jersey may also have a greater focus on quality measures by virtue of having been in quality programs such as Get With The Guidelines or by having been state designated and JC certified for a longer period of time.

The New Jersey requirements for CSCs, like those of the JC, include a large number of highly trained stroke experts, which ensures more continuous coverage. Although a disparity in mortality on weekends versus weekdays has been reported,[14] such a difference in mortality has not been seen at CSCs in New Jersey.[15] This lack of a weekend effect is felt to be related to the 24/7 availability of stroke specialists, advanced neuroimaging, ongoing training, and surveillance of specialized nursing care available at CSCs.[4, 16]

In our study, New Jersey CSCs overall had significantly higher rates of thrombolysis compared to PSCs (19.5% vs 9.6%) when looking at the 3‐hour window. This is higher overall than the national rate of 3.4% to 5.2%.[17] The number of patients treated in the expanded thrombolytic window were also significantly higher at CSCs, increasing thrombolysis rates to 22.4% at CSCs versus the 11% at PSCs. Door‐to‐drug times were also shorter at CSCs than PSCs in the 3‐ and 4.5‐hour windows (65 vs 74 minutes and 65 minutes vs 76 minutes, respectively). After we excluded transferred patients and those with a diagnosis of TIA, stroke not otherwise specified, and those with nonstroke‐related diagnoses, the total number of ischemic and hemorrhagic stroke patients seen at each of the 12 CSCs (n=11,505) was on average 4 times higher than the number seen at each of the 53 PSCs (n=12,732). High annual hospital stroke volume has been shown to be associated with higher rates of thrombolysis and lower stroke mortality.[14] A study of US academic centers found that although the same percentage of patients presented within 2 hours of stroke symptom onset in 2001 and 2004, the use of IV tPA more than doubled over this time period.[18] Improved system organization at the prehospital and hospital levels as well as greater comfort and experience with use of thrombolytic therapy likely contribute to all of these findings.[11]

CSCs did not outperform PSCs with regard to stroke education and antithrombotics by end of hospital day 2, but these results were not statistically significant. The former measure includes only stroke patients who are discharged home and is considered complete when all 5 of the following are addressed: risk factors for stroke, warning signs and symptoms for stroke, activation of emergency medical systems, follow‐up after discharge, and medications prescribed at discharge. CSCs were more likely to provide education for the first 3 and last component but less likely for the fourth element. These findings should be considered in the context of CSCs having higher volumes of more ill and complex patients who are more likely to be discharged to a rehabilitation hospital, nursing home, or other facility than to home. In our registry, CSCs discharged 46% of patients' to home versus 54% at PSCs. We speculate that CSCs may be less likely to habitually address follow‐up care and discharge medications as compared to PSCs. As far as provision of antithrombotics by hospital day 2, it is possible again that because CSCs have a higher number of complicated stroke patients, many may have had contraindications to use of antithrombotics in that time period.

Limitations of this study include the fact that this was a retrospective analysis of a database. Although the 2010 and 2011 NJASR dataset was sizeable, it was not possible to capture all potentially confounding variables that may have affected our point estimates. We were not able to perform a hierarchical analysis to account for clustering at the hospital level because of limited data available in the registry. Errors in recording data, coding, and documentation cannot be excluded. The fact that not all PSCs were necessarily JC certified may have contributed to the observed differences. Also, because pursuing PSC or CSC status is voluntary, it is not clear if the hospitals that chose CSC status were different in other unmeasured factors than those that chose PSC status, and the difference may have existed even in the absence of the designation program. Over the years, there have been changes in the criteria required by the state and the JC for PSC designation, although the larger differences between hospital levels remained intact. This may have limited our findings as well. The goal for hospitals is to continue strict adherence to policies and measures and thus improve quality of care for stroke patients. Future prospective studies should be conducted to ascertain validity and generalizability of our findings. Association of stroke measure adherence and functional outcomes would also be of interest. We were not able to measure this consistently in our study because not all patients at PSCs had admission and/or discharge NIHSS or modified Rankin Score. Although some studies have not shown an association between improved outcomes and higher performance on quality measures, we would like to look at this more closely in the stroke population.[19] As our database gets larger, we would like to reexamine our findings after correcting for more specific characteristics of each hospital. In the future, if additional states designate centers by level of stroke care, it will be important to compare how such designations compare to nonprofit organization certifications in terms of impacting performance on a larger scale.

CONCLUSION

This study shows better compliance of New Jersey state‐designated CSCs with the JC PSC core stroke measures and better mean door‐to‐thrombolytic drug times. Because these measures are evidence based, these results may translate into better stroke care and outcomes for patients treated at state‐designated CSCs.

Stroke is the fourth leading cause of death in the United States.[1] Though actual stroke‐related death has declined nationally by 19.4%, stroke‐related morbidity is still a significant burden.[2, 3] Hospital certification programs have been developed to improve the quality of stroke care on state and national levels. The Brain Attack Coalition (BAC) proposed 2 levels of stroke hospitals: primary stroke centers (PSCs) and comprehensive stroke centers (CSCs).[3, 4] Although most stroke patients can be cared for at PSCs, CSCs are able to care for more complex stroke patients.[4, 5] Using BAC recommendations the Joint Commission (JC) and American Heart Association/American Stroke Association created a certification program for PSCs. Eight evidence‐based performance measures are currently required for JC PSC certification.[6]

At the state level, New Jersey and Florida began designating PSCs and CSCs.[7, 8] New Jersey PSC and CSC designation criteria incorporate the elements of JC PSC certification, despite preceding them by several years.[6, 9] New Jersey CSC certification consists of more comprehensive requirements (Table 1). All New Jersey‐designated stroke centers submit data in the New Jersey Acute Stroke Registry (NJASR),[7] which closely matches the Centers for Disease Control and Prevention's Paul Coverdell National Acute Stroke Registry and includes the JC‐required core measures.

There is a paucity of data comparing state‐designated CSCs and PSCs, largely because few states have state designation programs. Although a recent observational study from Finland showed better outcomes in patients treated at CSCs, measures in that study were limited to mortality and institutionalization at 1 year.[10] In this study, we examined adherence of all New Jersey state‐designated stroke centers to the JC PSC measures and compared CSCs to PSCs in this regard. We posited that better compliance with these evidence‐based measures might translate into better quality of stroke care in the state and may lend support to future, larger studies that may be conducted because of the recent certification of CSCs by the JC.

METHODS

Components of the NJASR, key components of PSCs and CSCs in the BAC report, and the 8 JC's core stroke measures for PSC certification were assessed.[3, 4, 6]

First responders in New Jersey are required to bring suspected stroke patients to the nearest stroke‐designated hospital, regardless of whether it is a PSC or CSC, unless the patient is too medically unstable and needs to be taken to the nearest hospital. From there, decisions can be made to transfer a patient to a higher level hospital (CSC) depending on the complexity of the patient's condition.

New Jersey state‐designated PSCs and CSCs are required to abstract patient‐level data, evaluate outcomes, and initiate quality improvement activities on all patients evaluated for ischemic stroke, hemorrhagic stroke, transient ischemic attack (TIA), and those who undergo acute interventional therapy. Data are submitted quarterly to the New Jersey Department of Health and Senior Services (NJDHSS).[7] Hospital data are imported into the Acute Stroke Registry Database. The NJASR statewide dataset used for this analysis included all stroke admissions for the calendar years 2010 and 2011 and contains patient demographic information, health history, clinical investigations performed, treatments provided, and outcome measures that allow for risk‐adjusted assessment of outcomes.

The JC core stroke measures (STK‐1 thru 10 [except STK‐7 and 9]: venous thromboembolism [VTE] prophylaxis, discharged on antithrombotic therapy, anticoagulation therapy for atrial fibrillation/flutter, thrombolytic therapy, antithrombotic therapy by the end of hospital day 2, discharged on statin medication, stroke education, and assessment for rehabilitation) (except STK‐7 [dysphagia screening] and STK‐9 [smoking cessation/advice counseling]) apply only to acute ischemic and/or hemorrhagic stroke patients.

In our analysis, transferred patients and patients with a diagnosis of TIA, stroke not otherwise specified, and nonstroke‐related diagnoses were excluded. Hospital identity was kept anonymous through assignment of random numeric codes by the NJDHSS. Hospitals were categorized as CSC or PSC based on NJDHSS designation. Stroke severity on admission was assessed by categorizing National Institutes of Health Stroke Scale (NIHSS) scores into: no stroke when NIHSS=0, mild stroke when NIHSS=14, and moderate to severe stroke when NIHSS>5. Median door‐to‐thrombolytic drug times were assessed for both patients who arrived to the hospital within 2 hours of stroke symptom onset and received thrombolytic therapy within 3 hours, as well as patients who arrived within 3.5 hours (210 minutes) and received treatment within 4.5 hours (270 minutes).

RESULTS

There were 36,892 acute stroke cases treated at the 53 New Jersey PSCs and 12 CSCs in the calendar year 2010 and 2011 (Table 2). Sixty percent were treated at PSCs and 40% at CSCs. There were significant differences in the distribution of patients' characteristics (race, age, and gender) between the 2 hospital levels. At both PSCs and CSCs, the majority of patients were white, distantly followed by blacks. Patients at PSCs were statistically significantly older than CSCs. The most prevalent comorbid conditions in both PSCs and CSCs were hypertension, diabetes mellitus, and dyslipidemia. Based on our categorization, we found that 45% of patients admitted to CSCs had moderate‐to‐severe stroke (NIHSS>5). The median door‐to‐thrombolytic drug times were significantly shorter at CSCs than PSCs for both the 3‐hour (65 vs 74 minutes, P<0.0001) and 4.5 hour (65 vs 76 minutes, P<0.0001) IV tPA time windows.

The incidences of stroke diagnosis types are also detailed in Table 2. Seventy percent of patients at CSCs had either an ischemic or hemorrhagic stroke diagnosis versus 57.1% of patients admitted at PSCs. Hemorrhagic stroke patients were twice as likely to be admitted at CSCs compared to PSCs.

After excluding 13,964 patients with a diagnosis of TIA, stroke not otherwise specified, and those with nonstroke‐related diagnosis, the likelihood of stroke patients' receiving the JC's performance measure services at either of these hospital levels was assessed (Table 3). In general, the adjusted odds ratio estimates of patients receiving a JC core performance measure at PSCs were lower than CSCs, indicating better compliance with the measures at CSCs. For example, 19.5% of eligible patients received thrombolytic therapy at CSCs compared to 9.6% at PSCs. CSCs also were more likely to provide VTE prophylaxis, anticoagulation for atrial fibrillation, and assessment for rehabilitation. Stroke education and antithrombotic therapy by the end of hospital day 2 were more likely to be provided at PSCs, but the results were not statistically significant.

DISCUSSION

In New Jersey, CSCs were more likely to adhere better to JC core performance measures than PSCs. Median door‐to‐thrombolytic drug times were also significantly lower at CSCs. Such differences may be due to several factors including the fact that CSCs have generally been state designated for a longer period of time than PSCs. CSCs are likely to have higher volumes of stroke admissions, are more likely to be JC certified, provide more staff education, and have more staff and resources. The New Jersey stroke designation program began in 2006, and 11 of the 12 CSCs were designated by the end of 2007. However, the PSC designation process has been more gradual, with several of them designated in 2010 and 2011 as the data for this study were being collected.

The New York State Stroke Center Designation Project prospectively showed that stroke center designation improved the quality of acute stroke patient care and administration of thrombolytic therapy; however, differing levels of hospital designation were not present in New York at that time.[11] Participation in a data measurement program such as Get With The Guidelines has also been examined. It is evident that the amount of time in a program is predictive of process measure compliance.[12] JC certification as a PSC is also associated with increased thrombolytic rates for acute stroke over time.[13] New Jersey does not require that stroke‐designated hospitals have JC stroke certification. Although 11 New Jersey CSCs have been certified as JC PSCs since 2009, only 21 of the 53 state‐designated PSCs are JC certified. It may be that the highest performing sites pursue state CSC designation and JC PSC certification/recertification repeatedly. CSCs in New Jersey may also have a greater focus on quality measures by virtue of having been in quality programs such as Get With The Guidelines or by having been state designated and JC certified for a longer period of time.

The New Jersey requirements for CSCs, like those of the JC, include a large number of highly trained stroke experts, which ensures more continuous coverage. Although a disparity in mortality on weekends versus weekdays has been reported,[14] such a difference in mortality has not been seen at CSCs in New Jersey.[15] This lack of a weekend effect is felt to be related to the 24/7 availability of stroke specialists, advanced neuroimaging, ongoing training, and surveillance of specialized nursing care available at CSCs.[4, 16]

In our study, New Jersey CSCs overall had significantly higher rates of thrombolysis compared to PSCs (19.5% vs 9.6%) when looking at the 3‐hour window. This is higher overall than the national rate of 3.4% to 5.2%.[17] The number of patients treated in the expanded thrombolytic window were also significantly higher at CSCs, increasing thrombolysis rates to 22.4% at CSCs versus the 11% at PSCs. Door‐to‐drug times were also shorter at CSCs than PSCs in the 3‐ and 4.5‐hour windows (65 vs 74 minutes and 65 minutes vs 76 minutes, respectively). After we excluded transferred patients and those with a diagnosis of TIA, stroke not otherwise specified, and those with nonstroke‐related diagnoses, the total number of ischemic and hemorrhagic stroke patients seen at each of the 12 CSCs (n=11,505) was on average 4 times higher than the number seen at each of the 53 PSCs (n=12,732). High annual hospital stroke volume has been shown to be associated with higher rates of thrombolysis and lower stroke mortality.[14] A study of US academic centers found that although the same percentage of patients presented within 2 hours of stroke symptom onset in 2001 and 2004, the use of IV tPA more than doubled over this time period.[18] Improved system organization at the prehospital and hospital levels as well as greater comfort and experience with use of thrombolytic therapy likely contribute to all of these findings.[11]

CSCs did not outperform PSCs with regard to stroke education and antithrombotics by end of hospital day 2, but these results were not statistically significant. The former measure includes only stroke patients who are discharged home and is considered complete when all 5 of the following are addressed: risk factors for stroke, warning signs and symptoms for stroke, activation of emergency medical systems, follow‐up after discharge, and medications prescribed at discharge. CSCs were more likely to provide education for the first 3 and last component but less likely for the fourth element. These findings should be considered in the context of CSCs having higher volumes of more ill and complex patients who are more likely to be discharged to a rehabilitation hospital, nursing home, or other facility than to home. In our registry, CSCs discharged 46% of patients' to home versus 54% at PSCs. We speculate that CSCs may be less likely to habitually address follow‐up care and discharge medications as compared to PSCs. As far as provision of antithrombotics by hospital day 2, it is possible again that because CSCs have a higher number of complicated stroke patients, many may have had contraindications to use of antithrombotics in that time period.

Limitations of this study include the fact that this was a retrospective analysis of a database. Although the 2010 and 2011 NJASR dataset was sizeable, it was not possible to capture all potentially confounding variables that may have affected our point estimates. We were not able to perform a hierarchical analysis to account for clustering at the hospital level because of limited data available in the registry. Errors in recording data, coding, and documentation cannot be excluded. The fact that not all PSCs were necessarily JC certified may have contributed to the observed differences. Also, because pursuing PSC or CSC status is voluntary, it is not clear if the hospitals that chose CSC status were different in other unmeasured factors than those that chose PSC status, and the difference may have existed even in the absence of the designation program. Over the years, there have been changes in the criteria required by the state and the JC for PSC designation, although the larger differences between hospital levels remained intact. This may have limited our findings as well. The goal for hospitals is to continue strict adherence to policies and measures and thus improve quality of care for stroke patients. Future prospective studies should be conducted to ascertain validity and generalizability of our findings. Association of stroke measure adherence and functional outcomes would also be of interest. We were not able to measure this consistently in our study because not all patients at PSCs had admission and/or discharge NIHSS or modified Rankin Score. Although some studies have not shown an association between improved outcomes and higher performance on quality measures, we would like to look at this more closely in the stroke population.[19] As our database gets larger, we would like to reexamine our findings after correcting for more specific characteristics of each hospital. In the future, if additional states designate centers by level of stroke care, it will be important to compare how such designations compare to nonprofit organization certifications in terms of impacting performance on a larger scale.

CONCLUSION

This study shows better compliance of New Jersey state‐designated CSCs with the JC PSC core stroke measures and better mean door‐to‐thrombolytic drug times. Because these measures are evidence based, these results may translate into better stroke care and outcomes for patients treated at state‐designated CSCs.

Stroke is the fourth leading cause of death in the United States.[1] Though actual stroke‐related death has declined nationally by 19.4%, stroke‐related morbidity is still a significant burden.[2, 3] Hospital certification programs have been developed to improve the quality of stroke care on state and national levels. The Brain Attack Coalition (BAC) proposed 2 levels of stroke hospitals: primary stroke centers (PSCs) and comprehensive stroke centers (CSCs).[3, 4] Although most stroke patients can be cared for at PSCs, CSCs are able to care for more complex stroke patients.[4, 5] Using BAC recommendations the Joint Commission (JC) and American Heart Association/American Stroke Association created a certification program for PSCs. Eight evidence‐based performance measures are currently required for JC PSC certification.[6]

At the state level, New Jersey and Florida began designating PSCs and CSCs.[7, 8] New Jersey PSC and CSC designation criteria incorporate the elements of JC PSC certification, despite preceding them by several years.[6, 9] New Jersey CSC certification consists of more comprehensive requirements (Table 1). All New Jersey‐designated stroke centers submit data in the New Jersey Acute Stroke Registry (NJASR),[7] which closely matches the Centers for Disease Control and Prevention's Paul Coverdell National Acute Stroke Registry and includes the JC‐required core measures.

There is a paucity of data comparing state‐designated CSCs and PSCs, largely because few states have state designation programs. Although a recent observational study from Finland showed better outcomes in patients treated at CSCs, measures in that study were limited to mortality and institutionalization at 1 year.[10] In this study, we examined adherence of all New Jersey state‐designated stroke centers to the JC PSC measures and compared CSCs to PSCs in this regard. We posited that better compliance with these evidence‐based measures might translate into better quality of stroke care in the state and may lend support to future, larger studies that may be conducted because of the recent certification of CSCs by the JC.

METHODS

Components of the NJASR, key components of PSCs and CSCs in the BAC report, and the 8 JC's core stroke measures for PSC certification were assessed.[3, 4, 6]

First responders in New Jersey are required to bring suspected stroke patients to the nearest stroke‐designated hospital, regardless of whether it is a PSC or CSC, unless the patient is too medically unstable and needs to be taken to the nearest hospital. From there, decisions can be made to transfer a patient to a higher level hospital (CSC) depending on the complexity of the patient's condition.

New Jersey state‐designated PSCs and CSCs are required to abstract patient‐level data, evaluate outcomes, and initiate quality improvement activities on all patients evaluated for ischemic stroke, hemorrhagic stroke, transient ischemic attack (TIA), and those who undergo acute interventional therapy. Data are submitted quarterly to the New Jersey Department of Health and Senior Services (NJDHSS).[7] Hospital data are imported into the Acute Stroke Registry Database. The NJASR statewide dataset used for this analysis included all stroke admissions for the calendar years 2010 and 2011 and contains patient demographic information, health history, clinical investigations performed, treatments provided, and outcome measures that allow for risk‐adjusted assessment of outcomes.

The JC core stroke measures (STK‐1 thru 10 [except STK‐7 and 9]: venous thromboembolism [VTE] prophylaxis, discharged on antithrombotic therapy, anticoagulation therapy for atrial fibrillation/flutter, thrombolytic therapy, antithrombotic therapy by the end of hospital day 2, discharged on statin medication, stroke education, and assessment for rehabilitation) (except STK‐7 [dysphagia screening] and STK‐9 [smoking cessation/advice counseling]) apply only to acute ischemic and/or hemorrhagic stroke patients.

In our analysis, transferred patients and patients with a diagnosis of TIA, stroke not otherwise specified, and nonstroke‐related diagnoses were excluded. Hospital identity was kept anonymous through assignment of random numeric codes by the NJDHSS. Hospitals were categorized as CSC or PSC based on NJDHSS designation. Stroke severity on admission was assessed by categorizing National Institutes of Health Stroke Scale (NIHSS) scores into: no stroke when NIHSS=0, mild stroke when NIHSS=14, and moderate to severe stroke when NIHSS>5. Median door‐to‐thrombolytic drug times were assessed for both patients who arrived to the hospital within 2 hours of stroke symptom onset and received thrombolytic therapy within 3 hours, as well as patients who arrived within 3.5 hours (210 minutes) and received treatment within 4.5 hours (270 minutes).

RESULTS

There were 36,892 acute stroke cases treated at the 53 New Jersey PSCs and 12 CSCs in the calendar year 2010 and 2011 (Table 2). Sixty percent were treated at PSCs and 40% at CSCs. There were significant differences in the distribution of patients' characteristics (race, age, and gender) between the 2 hospital levels. At both PSCs and CSCs, the majority of patients were white, distantly followed by blacks. Patients at PSCs were statistically significantly older than CSCs. The most prevalent comorbid conditions in both PSCs and CSCs were hypertension, diabetes mellitus, and dyslipidemia. Based on our categorization, we found that 45% of patients admitted to CSCs had moderate‐to‐severe stroke (NIHSS>5). The median door‐to‐thrombolytic drug times were significantly shorter at CSCs than PSCs for both the 3‐hour (65 vs 74 minutes, P<0.0001) and 4.5 hour (65 vs 76 minutes, P<0.0001) IV tPA time windows.

The incidences of stroke diagnosis types are also detailed in Table 2. Seventy percent of patients at CSCs had either an ischemic or hemorrhagic stroke diagnosis versus 57.1% of patients admitted at PSCs. Hemorrhagic stroke patients were twice as likely to be admitted at CSCs compared to PSCs.

After excluding 13,964 patients with a diagnosis of TIA, stroke not otherwise specified, and those with nonstroke‐related diagnosis, the likelihood of stroke patients' receiving the JC's performance measure services at either of these hospital levels was assessed (Table 3). In general, the adjusted odds ratio estimates of patients receiving a JC core performance measure at PSCs were lower than CSCs, indicating better compliance with the measures at CSCs. For example, 19.5% of eligible patients received thrombolytic therapy at CSCs compared to 9.6% at PSCs. CSCs also were more likely to provide VTE prophylaxis, anticoagulation for atrial fibrillation, and assessment for rehabilitation. Stroke education and antithrombotic therapy by the end of hospital day 2 were more likely to be provided at PSCs, but the results were not statistically significant.

DISCUSSION

In New Jersey, CSCs were more likely to adhere better to JC core performance measures than PSCs. Median door‐to‐thrombolytic drug times were also significantly lower at CSCs. Such differences may be due to several factors including the fact that CSCs have generally been state designated for a longer period of time than PSCs. CSCs are likely to have higher volumes of stroke admissions, are more likely to be JC certified, provide more staff education, and have more staff and resources. The New Jersey stroke designation program began in 2006, and 11 of the 12 CSCs were designated by the end of 2007. However, the PSC designation process has been more gradual, with several of them designated in 2010 and 2011 as the data for this study were being collected.

The New York State Stroke Center Designation Project prospectively showed that stroke center designation improved the quality of acute stroke patient care and administration of thrombolytic therapy; however, differing levels of hospital designation were not present in New York at that time.[11] Participation in a data measurement program such as Get With The Guidelines has also been examined. It is evident that the amount of time in a program is predictive of process measure compliance.[12] JC certification as a PSC is also associated with increased thrombolytic rates for acute stroke over time.[13] New Jersey does not require that stroke‐designated hospitals have JC stroke certification. Although 11 New Jersey CSCs have been certified as JC PSCs since 2009, only 21 of the 53 state‐designated PSCs are JC certified. It may be that the highest performing sites pursue state CSC designation and JC PSC certification/recertification repeatedly. CSCs in New Jersey may also have a greater focus on quality measures by virtue of having been in quality programs such as Get With The Guidelines or by having been state designated and JC certified for a longer period of time.

The New Jersey requirements for CSCs, like those of the JC, include a large number of highly trained stroke experts, which ensures more continuous coverage. Although a disparity in mortality on weekends versus weekdays has been reported,[14] such a difference in mortality has not been seen at CSCs in New Jersey.[15] This lack of a weekend effect is felt to be related to the 24/7 availability of stroke specialists, advanced neuroimaging, ongoing training, and surveillance of specialized nursing care available at CSCs.[4, 16]

In our study, New Jersey CSCs overall had significantly higher rates of thrombolysis compared to PSCs (19.5% vs 9.6%) when looking at the 3‐hour window. This is higher overall than the national rate of 3.4% to 5.2%.[17] The number of patients treated in the expanded thrombolytic window were also significantly higher at CSCs, increasing thrombolysis rates to 22.4% at CSCs versus the 11% at PSCs. Door‐to‐drug times were also shorter at CSCs than PSCs in the 3‐ and 4.5‐hour windows (65 vs 74 minutes and 65 minutes vs 76 minutes, respectively). After we excluded transferred patients and those with a diagnosis of TIA, stroke not otherwise specified, and those with nonstroke‐related diagnoses, the total number of ischemic and hemorrhagic stroke patients seen at each of the 12 CSCs (n=11,505) was on average 4 times higher than the number seen at each of the 53 PSCs (n=12,732). High annual hospital stroke volume has been shown to be associated with higher rates of thrombolysis and lower stroke mortality.[14] A study of US academic centers found that although the same percentage of patients presented within 2 hours of stroke symptom onset in 2001 and 2004, the use of IV tPA more than doubled over this time period.[18] Improved system organization at the prehospital and hospital levels as well as greater comfort and experience with use of thrombolytic therapy likely contribute to all of these findings.[11]

CSCs did not outperform PSCs with regard to stroke education and antithrombotics by end of hospital day 2, but these results were not statistically significant. The former measure includes only stroke patients who are discharged home and is considered complete when all 5 of the following are addressed: risk factors for stroke, warning signs and symptoms for stroke, activation of emergency medical systems, follow‐up after discharge, and medications prescribed at discharge. CSCs were more likely to provide education for the first 3 and last component but less likely for the fourth element. These findings should be considered in the context of CSCs having higher volumes of more ill and complex patients who are more likely to be discharged to a rehabilitation hospital, nursing home, or other facility than to home. In our registry, CSCs discharged 46% of patients' to home versus 54% at PSCs. We speculate that CSCs may be less likely to habitually address follow‐up care and discharge medications as compared to PSCs. As far as provision of antithrombotics by hospital day 2, it is possible again that because CSCs have a higher number of complicated stroke patients, many may have had contraindications to use of antithrombotics in that time period.

Limitations of this study include the fact that this was a retrospective analysis of a database. Although the 2010 and 2011 NJASR dataset was sizeable, it was not possible to capture all potentially confounding variables that may have affected our point estimates. We were not able to perform a hierarchical analysis to account for clustering at the hospital level because of limited data available in the registry. Errors in recording data, coding, and documentation cannot be excluded. The fact that not all PSCs were necessarily JC certified may have contributed to the observed differences. Also, because pursuing PSC or CSC status is voluntary, it is not clear if the hospitals that chose CSC status were different in other unmeasured factors than those that chose PSC status, and the difference may have existed even in the absence of the designation program. Over the years, there have been changes in the criteria required by the state and the JC for PSC designation, although the larger differences between hospital levels remained intact. This may have limited our findings as well. The goal for hospitals is to continue strict adherence to policies and measures and thus improve quality of care for stroke patients. Future prospective studies should be conducted to ascertain validity and generalizability of our findings. Association of stroke measure adherence and functional outcomes would also be of interest. We were not able to measure this consistently in our study because not all patients at PSCs had admission and/or discharge NIHSS or modified Rankin Score. Although some studies have not shown an association between improved outcomes and higher performance on quality measures, we would like to look at this more closely in the stroke population.[19] As our database gets larger, we would like to reexamine our findings after correcting for more specific characteristics of each hospital. In the future, if additional states designate centers by level of stroke care, it will be important to compare how such designations compare to nonprofit organization certifications in terms of impacting performance on a larger scale.

CONCLUSION

This study shows better compliance of New Jersey state‐designated CSCs with the JC PSC core stroke measures and better mean door‐to‐thrombolytic drug times. Because these measures are evidence based, these results may translate into better stroke care and outcomes for patients treated at state‐designated CSCs.

Inpatient admissions are a major component of healthcare costs in the United States,[1] where the number of annual inpatient hospital admissions has increased by 15% from 34.3 million in 1993 to 39.5 million in 2006.² Studies performed predominantly in Europe have shown that inappropriate use of hospital beds exceeds 20% across various specialties.[3] A study by Campbell et al. showed that if given the choice, 60% of physicians would consider an alternative to admission for such patients, if such an option were available, and 70% of patients would prefer not to be admitted for workup.[4] Based on similar findings, various hospitals across the world have tried to make organizational changes to allocate healthcare resources more efficiently. The concept of quick and early diagnosis was first introduced in 1996 by Kendall et al., and it included a hospital unit in the United Kingdom managed by consultants receiving referrals from primary care doctors and led to early diagnostic workup without hospitalization.[5] A more refined version of this concept, a potentially cost‐saving and efficient alternative to inpatient hospitalization for diagnostic purposes, was described by Bosch et al., and named the quick diagnosis unit (QDU).[1]

The basic objectives of QDUs include early diagnosis of potentially severe diseases such as cancer, avoiding unnecessary hospitalization, minimizing hospital morbidity, reducing costs, and improving patient satisfaction. The first described QDU was managed by internists, where patients with specific symptoms such as undiagnosed lumps or masses, anemia, hematuria, or gastrointestinal symptoms could be referred for a diagnostic evaluation. Patients were required to be well enough to travel to the QDU on an outpatient basis, and patients unable to do so were thought to be better suited for hospitalization.[1]

In the present study, we conduct a systematic review, the first one on this subject to our knowledge, of studies that tested established QDUs or similar units in hospital settings. The majority of established units were tested and exist in Europe.[1, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14] They have been studied in Spain, from where much of these data have been obtained.[1]

METHODS

RESULTS

Study Selection

Our literature search initially yielded 2047 publications, out of which 2034 were excluded after title and abstract screening. Thirteen studies were selected for full‐text review, out of which 5 were selected for detailed review based on our inclusion criteria (Figure 1). Three of the studies were in Spanish, and the results were analyzed with the help of a Spanish translator. The other 2 studies were in English.

Figure 1

DISCUSSION

Our systematic review evaluated the effectiveness of QDUs for the diagnostic evaluation of patients with potentially severe disease and showed that such units, where established, are cost‐effective, prevent unnecessary hospitalizations, and diagnose potentially severe diseases, particularly malignant conditions, in a timely manner.

QDUs can evaluate medically stable patients with a variety of complaints such as anemia, lymphadenopathy, undiagnosed lumps and masses, and gastrointestinal symptoms and accelerate the diagnostic evaluation without requiring inpatient hospitalization. Many times patients are admitted to the hospital for a diagnostic evaluation without actual treatment.[12] These patients may not be sick enough to warrant hospitalization and may be able to return to the clinic for an outpatient workup. The QDU approach can complete the evaluation in such patients with the added advantages of saving money and higher patient satisfaction, due to diminished disruption of the patient's daily life.[1, 12] As most primary care physicians are unlikely to provide regular and frequent access for unscheduled care and EDs are more likely to admit patients for diagnostic workup,[2] a QDU approach seems a reasonable alternative for making a quick diagnosis and at the same time avoiding unnecessary hospitalization. Bosch et al. have also evaluated the impact of the QDUs in the diagnosis of specific diseases such as cancer in 169 patients diagnosed at the QDU, and compared them to 53 patients who were diagnosed with cancer during an inpatient evaluation.[11] They found that although QDU patients were significantly younger than hospitalized patients, there was no difference in diagnoses established and the time to diagnosis at the QDU and length of stay in the hospital.

There is a significant cost saving associated with QDUs. The cost savings calculated by Bosch et al. and Capell et al. were for each patient enrolled in this protocol from index encounter to final diagnosis.[6, 7] These 2 studies describe primarily fixed costs saved per patient treated in the QDU versus an inpatient admission. Fixed costs in hospital care include personnel cost, buildings, and equipment, whereas variable costs include medication, test reagents, and disposable supplies.[15] In comparison with the US healthcare system, fixed costs in Europe are considerably lower, and certain variable costs (like medications and procedures) are significantly higher in the United States.[16] This suggests a greater opportunity for healthcare savings for carefully selected patients in the United States, where costs related to inpatient admissions are significantly higher.[16]

Another limitation of our analysis is the paucity of studies on this topic. Many of the publications are from Bosch et al.,[1, 6, 11, 12, 13, 14] a single group in Spain, and these show considerable cost savings, patient satisfaction, and patient safety. However, most of their data are either retrospective or from nonrandomized, prospective cohort studies. The only report describing a similar approach in the United States was by Paschal in the city of New Orleans.[17] After Charity Hospital and the Veterans Affairs Hospital in New Orleans were lost to hurricane Katrina, an urgent care clinic was set up where potentially severe diseases such as cancer, leukemia, and autoimmune and endocrine disorders were diagnosed efficiently, although safety data were not reported.

The reported studies used different study designs and evaluated different primary outcomes. These limitations can be overcome with a well‐designed prospective trial, which could also evaluate the actual impact on patient care, safety, and healthcare savings in the United States.

Safety data were reported in detail only in 1 study,[6] and the rates of admissions were reported by 2 other studies.[7, 8] These suggest that QDUs may be safe for a selected group of patients. Patients evaluated in these units preferred this approach as shown by the overwhelming majority of the patients who chose QDU care over inpatient admissions when patient surveys were performed.

CONCLUSION

In this era of healthcare reform and emphasis on value‐based care, we must optimize the efficiency of our care delivery systems and challenge our preexisting resource‐intensive healthcare models. One source of potential savings is avoiding hospitalizations for purely diagnostic purposes, utilizing quick diagnostic units for patients who are able to return for outpatient evaluations. Such units are established, have been studied in Europe, and our systematic review shows that they are cost‐effective, time‐ and resource‐efficient, and preferred by patients. In our healthcare system, with the high cost of inpatient care, the QDU can yield large savings of healthcare dollars while expediting diagnostic workup, increasing patient satisfaction, and preventing lost productivity from hospital stays. Further exploration and study of alternative care delivery models, such as quick diagnostic units, is required to achieve the goal of cost‐effective high‐quality care for all.

Inpatient admissions are a major component of healthcare costs in the United States,[1] where the number of annual inpatient hospital admissions has increased by 15% from 34.3 million in 1993 to 39.5 million in 2006.² Studies performed predominantly in Europe have shown that inappropriate use of hospital beds exceeds 20% across various specialties.[3] A study by Campbell et al. showed that if given the choice, 60% of physicians would consider an alternative to admission for such patients, if such an option were available, and 70% of patients would prefer not to be admitted for workup.[4] Based on similar findings, various hospitals across the world have tried to make organizational changes to allocate healthcare resources more efficiently. The concept of quick and early diagnosis was first introduced in 1996 by Kendall et al., and it included a hospital unit in the United Kingdom managed by consultants receiving referrals from primary care doctors and led to early diagnostic workup without hospitalization.[5] A more refined version of this concept, a potentially cost‐saving and efficient alternative to inpatient hospitalization for diagnostic purposes, was described by Bosch et al., and named the quick diagnosis unit (QDU).[1]

The basic objectives of QDUs include early diagnosis of potentially severe diseases such as cancer, avoiding unnecessary hospitalization, minimizing hospital morbidity, reducing costs, and improving patient satisfaction. The first described QDU was managed by internists, where patients with specific symptoms such as undiagnosed lumps or masses, anemia, hematuria, or gastrointestinal symptoms could be referred for a diagnostic evaluation. Patients were required to be well enough to travel to the QDU on an outpatient basis, and patients unable to do so were thought to be better suited for hospitalization.[1]

In the present study, we conduct a systematic review, the first one on this subject to our knowledge, of studies that tested established QDUs or similar units in hospital settings. The majority of established units were tested and exist in Europe.[1, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14] They have been studied in Spain, from where much of these data have been obtained.[1]

METHODS

RESULTS

Study Selection

Our literature search initially yielded 2047 publications, out of which 2034 were excluded after title and abstract screening. Thirteen studies were selected for full‐text review, out of which 5 were selected for detailed review based on our inclusion criteria (Figure 1). Three of the studies were in Spanish, and the results were analyzed with the help of a Spanish translator. The other 2 studies were in English.

Figure 1

DISCUSSION

Our systematic review evaluated the effectiveness of QDUs for the diagnostic evaluation of patients with potentially severe disease and showed that such units, where established, are cost‐effective, prevent unnecessary hospitalizations, and diagnose potentially severe diseases, particularly malignant conditions, in a timely manner.

QDUs can evaluate medically stable patients with a variety of complaints such as anemia, lymphadenopathy, undiagnosed lumps and masses, and gastrointestinal symptoms and accelerate the diagnostic evaluation without requiring inpatient hospitalization. Many times patients are admitted to the hospital for a diagnostic evaluation without actual treatment.[12] These patients may not be sick enough to warrant hospitalization and may be able to return to the clinic for an outpatient workup. The QDU approach can complete the evaluation in such patients with the added advantages of saving money and higher patient satisfaction, due to diminished disruption of the patient's daily life.[1, 12] As most primary care physicians are unlikely to provide regular and frequent access for unscheduled care and EDs are more likely to admit patients for diagnostic workup,[2] a QDU approach seems a reasonable alternative for making a quick diagnosis and at the same time avoiding unnecessary hospitalization. Bosch et al. have also evaluated the impact of the QDUs in the diagnosis of specific diseases such as cancer in 169 patients diagnosed at the QDU, and compared them to 53 patients who were diagnosed with cancer during an inpatient evaluation.[11] They found that although QDU patients were significantly younger than hospitalized patients, there was no difference in diagnoses established and the time to diagnosis at the QDU and length of stay in the hospital.

There is a significant cost saving associated with QDUs. The cost savings calculated by Bosch et al. and Capell et al. were for each patient enrolled in this protocol from index encounter to final diagnosis.[6, 7] These 2 studies describe primarily fixed costs saved per patient treated in the QDU versus an inpatient admission. Fixed costs in hospital care include personnel cost, buildings, and equipment, whereas variable costs include medication, test reagents, and disposable supplies.[15] In comparison with the US healthcare system, fixed costs in Europe are considerably lower, and certain variable costs (like medications and procedures) are significantly higher in the United States.[16] This suggests a greater opportunity for healthcare savings for carefully selected patients in the United States, where costs related to inpatient admissions are significantly higher.[16]

Another limitation of our analysis is the paucity of studies on this topic. Many of the publications are from Bosch et al.,[1, 6, 11, 12, 13, 14] a single group in Spain, and these show considerable cost savings, patient satisfaction, and patient safety. However, most of their data are either retrospective or from nonrandomized, prospective cohort studies. The only report describing a similar approach in the United States was by Paschal in the city of New Orleans.[17] After Charity Hospital and the Veterans Affairs Hospital in New Orleans were lost to hurricane Katrina, an urgent care clinic was set up where potentially severe diseases such as cancer, leukemia, and autoimmune and endocrine disorders were diagnosed efficiently, although safety data were not reported.

The reported studies used different study designs and evaluated different primary outcomes. These limitations can be overcome with a well‐designed prospective trial, which could also evaluate the actual impact on patient care, safety, and healthcare savings in the United States.

Safety data were reported in detail only in 1 study,[6] and the rates of admissions were reported by 2 other studies.[7, 8] These suggest that QDUs may be safe for a selected group of patients. Patients evaluated in these units preferred this approach as shown by the overwhelming majority of the patients who chose QDU care over inpatient admissions when patient surveys were performed.

CONCLUSION

In this era of healthcare reform and emphasis on value‐based care, we must optimize the efficiency of our care delivery systems and challenge our preexisting resource‐intensive healthcare models. One source of potential savings is avoiding hospitalizations for purely diagnostic purposes, utilizing quick diagnostic units for patients who are able to return for outpatient evaluations. Such units are established, have been studied in Europe, and our systematic review shows that they are cost‐effective, time‐ and resource‐efficient, and preferred by patients. In our healthcare system, with the high cost of inpatient care, the QDU can yield large savings of healthcare dollars while expediting diagnostic workup, increasing patient satisfaction, and preventing lost productivity from hospital stays. Further exploration and study of alternative care delivery models, such as quick diagnostic units, is required to achieve the goal of cost‐effective high‐quality care for all.

Inpatient admissions are a major component of healthcare costs in the United States,[1] where the number of annual inpatient hospital admissions has increased by 15% from 34.3 million in 1993 to 39.5 million in 2006.² Studies performed predominantly in Europe have shown that inappropriate use of hospital beds exceeds 20% across various specialties.[3] A study by Campbell et al. showed that if given the choice, 60% of physicians would consider an alternative to admission for such patients, if such an option were available, and 70% of patients would prefer not to be admitted for workup.[4] Based on similar findings, various hospitals across the world have tried to make organizational changes to allocate healthcare resources more efficiently. The concept of quick and early diagnosis was first introduced in 1996 by Kendall et al., and it included a hospital unit in the United Kingdom managed by consultants receiving referrals from primary care doctors and led to early diagnostic workup without hospitalization.[5] A more refined version of this concept, a potentially cost‐saving and efficient alternative to inpatient hospitalization for diagnostic purposes, was described by Bosch et al., and named the quick diagnosis unit (QDU).[1]

The basic objectives of QDUs include early diagnosis of potentially severe diseases such as cancer, avoiding unnecessary hospitalization, minimizing hospital morbidity, reducing costs, and improving patient satisfaction. The first described QDU was managed by internists, where patients with specific symptoms such as undiagnosed lumps or masses, anemia, hematuria, or gastrointestinal symptoms could be referred for a diagnostic evaluation. Patients were required to be well enough to travel to the QDU on an outpatient basis, and patients unable to do so were thought to be better suited for hospitalization.[1]

In the present study, we conduct a systematic review, the first one on this subject to our knowledge, of studies that tested established QDUs or similar units in hospital settings. The majority of established units were tested and exist in Europe.[1, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14] They have been studied in Spain, from where much of these data have been obtained.[1]

METHODS

RESULTS

Study Selection

Our literature search initially yielded 2047 publications, out of which 2034 were excluded after title and abstract screening. Thirteen studies were selected for full‐text review, out of which 5 were selected for detailed review based on our inclusion criteria (Figure 1). Three of the studies were in Spanish, and the results were analyzed with the help of a Spanish translator. The other 2 studies were in English.

Figure 1

DISCUSSION

Our systematic review evaluated the effectiveness of QDUs for the diagnostic evaluation of patients with potentially severe disease and showed that such units, where established, are cost‐effective, prevent unnecessary hospitalizations, and diagnose potentially severe diseases, particularly malignant conditions, in a timely manner.

QDUs can evaluate medically stable patients with a variety of complaints such as anemia, lymphadenopathy, undiagnosed lumps and masses, and gastrointestinal symptoms and accelerate the diagnostic evaluation without requiring inpatient hospitalization. Many times patients are admitted to the hospital for a diagnostic evaluation without actual treatment.[12] These patients may not be sick enough to warrant hospitalization and may be able to return to the clinic for an outpatient workup. The QDU approach can complete the evaluation in such patients with the added advantages of saving money and higher patient satisfaction, due to diminished disruption of the patient's daily life.[1, 12] As most primary care physicians are unlikely to provide regular and frequent access for unscheduled care and EDs are more likely to admit patients for diagnostic workup,[2] a QDU approach seems a reasonable alternative for making a quick diagnosis and at the same time avoiding unnecessary hospitalization. Bosch et al. have also evaluated the impact of the QDUs in the diagnosis of specific diseases such as cancer in 169 patients diagnosed at the QDU, and compared them to 53 patients who were diagnosed with cancer during an inpatient evaluation.[11] They found that although QDU patients were significantly younger than hospitalized patients, there was no difference in diagnoses established and the time to diagnosis at the QDU and length of stay in the hospital.

There is a significant cost saving associated with QDUs. The cost savings calculated by Bosch et al. and Capell et al. were for each patient enrolled in this protocol from index encounter to final diagnosis.[6, 7] These 2 studies describe primarily fixed costs saved per patient treated in the QDU versus an inpatient admission. Fixed costs in hospital care include personnel cost, buildings, and equipment, whereas variable costs include medication, test reagents, and disposable supplies.[15] In comparison with the US healthcare system, fixed costs in Europe are considerably lower, and certain variable costs (like medications and procedures) are significantly higher in the United States.[16] This suggests a greater opportunity for healthcare savings for carefully selected patients in the United States, where costs related to inpatient admissions are significantly higher.[16]

Another limitation of our analysis is the paucity of studies on this topic. Many of the publications are from Bosch et al.,[1, 6, 11, 12, 13, 14] a single group in Spain, and these show considerable cost savings, patient satisfaction, and patient safety. However, most of their data are either retrospective or from nonrandomized, prospective cohort studies. The only report describing a similar approach in the United States was by Paschal in the city of New Orleans.[17] After Charity Hospital and the Veterans Affairs Hospital in New Orleans were lost to hurricane Katrina, an urgent care clinic was set up where potentially severe diseases such as cancer, leukemia, and autoimmune and endocrine disorders were diagnosed efficiently, although safety data were not reported.

The reported studies used different study designs and evaluated different primary outcomes. These limitations can be overcome with a well‐designed prospective trial, which could also evaluate the actual impact on patient care, safety, and healthcare savings in the United States.

Safety data were reported in detail only in 1 study,[6] and the rates of admissions were reported by 2 other studies.[7, 8] These suggest that QDUs may be safe for a selected group of patients. Patients evaluated in these units preferred this approach as shown by the overwhelming majority of the patients who chose QDU care over inpatient admissions when patient surveys were performed.

CONCLUSION

In this era of healthcare reform and emphasis on value‐based care, we must optimize the efficiency of our care delivery systems and challenge our preexisting resource‐intensive healthcare models. One source of potential savings is avoiding hospitalizations for purely diagnostic purposes, utilizing quick diagnostic units for patients who are able to return for outpatient evaluations. Such units are established, have been studied in Europe, and our systematic review shows that they are cost‐effective, time‐ and resource‐efficient, and preferred by patients. In our healthcare system, with the high cost of inpatient care, the QDU can yield large savings of healthcare dollars while expediting diagnostic workup, increasing patient satisfaction, and preventing lost productivity from hospital stays. Further exploration and study of alternative care delivery models, such as quick diagnostic units, is required to achieve the goal of cost‐effective high‐quality care for all.