Outpatient Practice Patterns After Stroke Hospitalization Among Neurologists
Background and Purpose— Care after stroke hospitalization can provide several opportunities to optimize vascular risk reduction. However, not much is known about poststroke practice patterns among neurologists. Such knowledge may help direct specific efforts to improve the impact of practicing neurologists on clinical outcomes after stroke.
Methods— A survey soliciting information on processes of care in the outpatient setting after recent hospitalization for ischemic stroke or transient ischemic attack was mailed to a random sample of 833 US and Canadian neurologist–members of the American Academy of Neurology.
Results— A total of 475 (57%) responses were received. Practice demographics of survey responders and nonresponders were largely similar. Fourteen percent of respondents identified themselves as vascular neurologists. Overall, respondents reported frequently checking for medication adherence and counseling patients on lifestyle modification. However, neurologists reported screening more frequently for diabetes, hypertension, and dyslipidemia than actually treating these conditions (all P<0.0001) Vascular neurologists were more likely than general neurologists to screen for hypertension (97% versus 86%, P=0.016), dyslipidemia (94% versus 68%, P<0.001), diabetes (89% versus 62%, P<0.001), and sleep apnea (94% versus 79%, P=0.007) as well as to treat hypertension (71% versus 45%, P<0.001), dyslipidemia (82% versus 50%, P<0.001), diabetes (45% versus 21%, P<0.001), and current smoking (77% versus 59%, P=0.005). Neurologists with mostly government-insured and uninsured patients were significantly more likely to engage in vascular risk reduction treatment than neurologists with mostly commercially insured patients.
Conclusions— Self-reported rates of screening and treatment of major vascular risk factors by most neurologists after stroke hospitalization are substantial but not universal. Bridging knowledge gaps or adopting a systematic management approach in coordination with primary care physicians could help optimize poststroke care.
Vascular hazard in the first year after an incident transient ischemic attack (TIA) or ischemic stroke is substantial, with the highest recurrent vascular event risk occurring within 3 months of the index event.1,2 Fortunately, evidence-based data demonstrate that specific secondary preventive strategies can reduce vascular risk and improve clinical outcomes in high-risk populations, including those who have experienced a stroke. However, available evidence indicates that there is a gap between existing stroke prevention practices and those supported by the results of clinical trials.3–5 Specifically, there can be a high rate of discontinuation of secondary prevention therapies within months of discharge from the hospital.5 Such trends tragically and unnecessarily expose the patient with stroke or TIA to untoward risk during this time period of greatest risk for recurrence.6
Although neurologists often play a prominent role in the inpatient management of patients with stroke and TIA7 and are at the forefront of designing and optimizing standardized stroke inpatient pathways,8–10 we are unaware of any data reflecting the potential impact of practicing neurologists on the process and outcome of care of patients with ischemic stroke discharged from the hospital.7 The identification of variations in practice may provide insights into current poststroke hospitalization management trends among neurologists and help direct specific efforts geared at improving overall stroke patient outcomes.7 The objective of this study was to survey neurologists involved in the care of hospitalized patients with ischemic stroke and TIA and obtain information on the their poststroke hospitalization practice patterns.
Survey Instrument Development
Two of the authors (B.O. and J.S.) designed the first draft of the survey instrument incorporating recommended vascular risk reduction strategies per national consensus guidelines11 and key elements of poststroke hospitalization care used in a local secondary stroke prevention quality improvement project.10,12 The design, program tools, and favorable impact of this program on poststroke hospitalization treatment rates have been previously published.10,12 This initial draft and the goals of the study were reviewed and approved by members of the Stroke and Vascular Neurology Section of the American Academy of Neurology (AAN). Additional input from the AAN Stroke and Vascular Neurology Section, AAN Member Demographics Subcommittee, and AAN Surveys staff led to further refinement of the instrument. The instrument was then pilot-tested in December 2006 with Executive Committee members of the AAN Stroke and Vascular Neurology and General Neurology sections. The resulting participant (n=11) feedback was used to create a final consensus version of the survey instrument. This final self-administered 2-page instrument consisted of 13 items designed for completion in no longer than 15 minutes. The survey’s questions solicited information from participants regarding their practice demographics (eg, estimated number of patients with stroke cared for in a typical week, practice type, and so on), screening and prescription habits, frequency of patient stroke education, and use of serological tests when indicated.
A random sample from the AAN member database was drawn for the survey. The inclusion criteria for the population from which the sample was drawn were: having indicated cerebrovascular disease as a practice interest, having a US or Canadian mailing address, being a board-certified or board-eligible physician, and spending at least 25% professional time in clinical practice. Residents, fellows, retired physicians, and individuals who had received 3 AAN surveys in the preceding 3 years or who participated in the review or development of the survey instrument were excluded. The latter group included Executive Committee members of the AAN Stroke and Vascular Neurology and General Neurology sections as well as the Member Demographics Subcommittee. A total of 2103 members from the entire AAN membership roster matched all the prespecified study parameters. Eight hundred fifty members were randomly selected to be survey participants. Only 23 members from Canada were in the sample.
The survey instrument with a cover letter explaining its purpose signed by the Chair of the AAN Stroke and Vascular Neurology Section was first sent to the participant sample on February 5, 2007, by mail or fax. Participants were promised an incentive for participating, a $25 gift certificate to the AAN online store. All potential survey participants also received an e-mail with a link to the online version of the survey. Surveys returned as undeliverable were resent if alternative contact information was identified. A total of 3 reminders to complete the survey were sent to survey participants at approximately 3-week intervals. Data collection was closed on April 9, 2007. During the actual data collection, 17 participants were removed from the sample, the overwhelming majority of who specified that the survey content was not applicable to their current professional activities. This action left a final sample size of 833.
A crosscheck was performed to ensure there were no duplicates of names in the sample. To ensure that the number of responses had an absolute error of ±5% at the 95% CI, we needed feedback from 326 persons in the sample, ie, a 38% response rate. Responses were characterized by frequency distributions and standard descriptive statistics. Comparison of demographic and practice characteristics between respondents and nonrespondents was performed using χ2 analysis as well as t test and Wilcoxon signed rank test where appropriate. T o simplify the presentation of the analyses, responses to survey questions on screening and treatment of vascular risk factors were collapsed into 2 categories: screen (response options of “always,” “often,” “sometimes,” and “seldom”) versus do not screen (response options of “never screen” and “defer to primary care physician”).
The overall final response rate was 57% (475 respondents). The margin of error for all respondents at a 95% CI was ±4.5%. A comparison of profiles of survey respondents versus nonrespondents can be seen in Table 1. There were no significant differences between survey responders and nonresponders on the following demographic variables: age, gender, country of residence (US versus Canada), and professional time spent on clinical practice, teaching, administration, or other activities. However, significant differences were found between survey responders and nonresponders on the time spent in research. Significant differences were also found in the types of members who returned the questionnaire. Means for the survey items that solicited information on screening and management of stroke risk factors can be found in the Figure. Means closer to 6 indicated that members were engaging in the activities more often, whereas means closer to 1 indicated that members were deferring procedures to the primary care provider or doing them less often. Frequency of checking for patient adherence to antithrombotic and lipid modifier medications as well as counseling patients on lifestyle modification (risk of smoking, good dietary habits, regular exercise, and personal risk factors) by neurologists was fairly high (ie, mean >5 or often). On the other hand, actual management of diabetes, hypertension, dyslipidemia, and current smoking was relatively low (ie, mean, 1.9 to 3.4, never to seldom).
Neurologists reported screening their patients with stroke and TIA more than actually treating them for the following conditions: hypertension (z score=−14.89, P<0.0001), dyslipidemia (z score=−9.66, P<0.0001), and diabetes (z score=−13.11, P<0.0001). Patient insurance type did not have any significant influence on timing and frequency of follow-ups, screening for risk factors and medication adherence, or lifestyle modification counseling. However, neurologists who reported seeing mostly government-insured and uninsured patients compared with neurologists whose patients mostly had commercial insurance were more likely to engage in treatment of hypertension (53.6% versus 41.1%, P<0.018), diabetes (29% versus 18.5%, P<0.021), and dyslipidemia (58.6% versus 47.9%, P<0.042).
Table 2 displays a comparison of practice patterns between vascular versus general neurologists. General neurologists were significantly more likely to see a patient sooner for a follow-up appointment compared with vascular neurologists. There were no statistical differences between vascular and general neurologists in the frequency of follow-up visits. Compared with general neurologists, vascular neurologists were significantly more likely to screen for hypertension, dyslipidemia, diabetes, elevated liver functions, serum electrolytes, and obstructive sleep apnea. There were no differences between vascular and general neurologists in checking for adherence to antihypertensive and diabetes medications. Vascular neurologists were also significantly more likely to treat hypertension, dyslipidemia, diabetes, and smoking when compared with general neurologists.
Several studies have examined or are examining the late stage of the ischemic stroke patient management continuum by assessing longer-term patient adherence to prevention strategies after hospital discharge, but information on physician behavior during this period is lacking.5,13–15 The results of our survey of poststroke hospitalization practice patterns among AAN members revealed encouraging aspects, but also indicated that there are several areas in which neurologists might play a greater role in optimizing poststroke care. First, it was gratifying to observe the very high self-reported rates of lifestyle modification counseling and education about stroke by the neurologists surveyed. Although the survey did not seek to inquire whether the specific lifestyle instructions being conveyed by these physicians were in line with evidence-based studies and consensus guideline recommendations,11 generally it can be difficult to change human behavior.12 As such, regular and routine behavioral education of any kind can only serve to continually link the behavior with the recent ischemic cerebrovascular event and reinforce the need to adhere to healthy and knowledgeable lifestyle.12
Checking for medication adherence is also an important aspect of modifying poststroke vascular risk, particularly in light of several studies that have shown an unacceptably high rate of discontinuation of drugs later in the community after hospitalization for a vascular event.7 One particular study showed that only approximately two thirds of patients with stroke and TIA discharged from the hospital were on an antithrombotic agent 6 months later, and the leading reasons for discontinuation (excluding valid contraindications) were physician-related5 Other studies, which have shown relatively low rates of antithrombotic treatment among stroke survivors living in a long-term care facility, and suboptimal use of lipid modifiers in recent stroke survivors, suggest that maintaining appropriate drug treatment after stroke may be a widespread challenge.13,14 However, our survey respondents reported being adept at checking for adherence to the various vascular risk-reducing drugs commonly used in patients with stroke and TIA, which is also commendable.
On the other hand, when it came to the actual treatment of vascular conditions such as diabetes, hypertension, dyslipidemia, and current smoking, the majority of survey respondents never or rarely did this, with such treatment being left to the primary care physician. The brief survey was not designed to elicit whether stroke risk reduction differed depending on the role of neurologists. A variety of conditions might account for the disparity between checking for medication adherence versus initiating or modifying medications. These might include appropriate reliance on a very good primary care physician and the confusion that might result from 2 prescribers. A more intensive study would be necessary to know whether active prescribing by neurologists in the poststroke visit would result in better stroke risk reduction. In addition, we did not document whether lack of comfort with initiating and monitoring modern medications to treat the respective vascular risk factors was at play. If so, this might be a worthy target of educational programs. The significantly higher screening and treatment rates of vascular risk factors reported by vascular neurologists compared with general neurologists suggests that such educational programs could result in increased treatment rates by general neurologists. These observations were not related to a dearth of patients with stroke treated by general neurologists because all neurologists in the population surveyed indicated that at least one fourth of their practice involved managing cerebrovascular disease.
The lower rate of diabetes treatment compared with the treatment of other vascular risk factors has to be noted. This treatment difference may suggest that factors other than relying on primary care physician care could also explain the lack of neurologist involvement. However, this risk factor treatment disparity might also reflect the relative lack of prevention studies addressing diabetes in patients with stroke so far, as well as the absence of a current indication for an oral hypoglycemic agent in normoglycemic individuals, which is quite unlike the situation for antihypertensive use in normotensives or statin use in patients with near-normal low-density lipoprotein levels.11
Strong arguments can be made for neurologists aggressively treating risk factors in the poststroke patients. Patients may not have their risk factors adjusted in a timely fashion if they are unable to see primary care physicians promptly, like for example among patients who have not yet been allocated primary care physicians (particularly if they did not have one before the event). Even among patients with secure primary care physician relationships, the neurologist, as the disease expert in stroke and its risk factors, is in a special position to engage the poststroke patient to comply with medications to prevent second stroke; and the pathophysiological mechanism underlying most ischemic strokes, atherosclerosis, is a chronic progressive disease whose natural history can be greatly modified by continuous long-term treatment,11 thereby necessitating timely, regular, and aggressive intervention, not unlike the approach generally taken by neurologists in the management of epilepsy.
Bridging knowledge gaps and discrepancies between knowledge and actual practice may necessitate different approaches.7 One strategy could be practical case-based participatory educational sessions incorporating the current AAN-endorsed national stroke prevention guidelines11 presented at the annual AAN meeting or in the form of audiovisual materials. Another approach, which has already been successful in the inpatient stroke setting, would be to develop standardized poststroke outpatient algorithms such as a “stay with the guidelines” strategy with appropriate protocols and systems processes. Coordination with primary care givers is of course essential. The postdischarge component of the UCLA Preventing Recurrence of Thromboembolic Events through Coordinated Treatment (PROTECT) program was associated with relatively high treatment adherence rates and compliance with recommended biomarker goals at 3 months and 1 year.16
Another noteworthy result of our survey was the higher frequency of treating hypertension, diabetes, and dyslipidemia among neurologists mostly caring for patients with government insurance or no insurance compared with neurologists with mostly commercially insured patients. It is not immediately clear why this disparity may exist, but we can speculate that perhaps there is a real or presumed notion that the commercially insured patients are, or will be, promptly seen and managed by their primary care physicians without the bottlenecks that may hamper management of patients with government or no insurance.
This study has some limitations. These results were based on self-report and it is well known that discrepancies may exist between self-report and actual practices by physicians. However, these discrepancies are often due to physicians reporting higher rates of screening or treatment when compared with patient survey and chart audit.17,18 As such, our survey results may actually be underestimating the magnitude of the lower screening and treatment rates of patients with stroke and TIA postdischarge by neurologists. Furthermore, we excluded neurologists who were not devoting at least 25% of their time to managing patients with cerebrovascular disease. We suspect that screening and treatment rates among these physicians may likely be less than that found among the neurologists we sampled. Lastly, we did not require a strict definition, including fellowship training or board certification, for those who identified themselves as vascular neurologists. Despite the aforementioned limitations, our results highlight potential opportunities to improve the neurologist’s role in poststroke management. Because prevention remains our best option for reducing the societal burden from stroke,19 the findings can serve as an important first step toward exploring ways to better improve the care that we neurologists administer to our patients with stroke and TIA.
Lawrence W. Brass, MD, wholeheartedly supported this project and helped bring it to the AAN Stroke and Vascular Neurology Section for consideration. We thank the AAN surveys staff for distributing the survey, sending reminders, keeping track of responses, and conducting statistical analyses. We also thank the members of the Stroke and Vascular Neurology Section, General Neurology Section, and Member Demographics Subcommittee of the AAN for their support and constructive contributions to the success of this project. Finally, we appreciate the time taken by our many colleagues in the AAN to participate in this survey despite their busy schedules.
Bruce Ovbiagele has served as a scientific consultant to Boehringer Ingelheim and Bayer. He is on a Speaker’s Bureau for Boehringer Ingelheim, Sanofi-Aventis, and Bristol-Myers-Squibb.
Jeffrey Saver has served as a scientific consultant to Boehringer Ingelheim and Pfizer. He is on a Speaker’s Bureau for Boehringer Ingelheim.
- Received September 18, 2007.
- Revision received October 29, 2007.
- Accepted November 2, 2007.
Coull A, Lovett JK, Rothwell PM. Population based study of early risk of stroke after transient ischaemic attack or minor stroke: implications for public education and organisation of services. BMJ. 2004; 328: 326.
Mouradian M, Majumdar SR, Senthilselvan A, Khan K, Shuaib A. How well are hypertension, hyperlipidemia, diabetes, and smoking managed after a stroke or transient ischemic attack? Stroke. 2002; 33: 1656–1659.
Hillen T, Dundas R, Lawrence E, Stewart JA, Rudd AG, Wolfe CD. Antithrombotic and antihypertensive management 3 months after ischemic stroke: a prospective study in an inner city population. Stroke. 2000; 31: 469–475.
Dergalust S, Rofail MT, Bateshansky DS, Cohen SN. The incidence of discontinuation of extended release dipyridamole/aspirin, clopidogrel and warfarin for the secondary prevention of stroke. Presented at the 55th American Academy of Neurology Annual Meeting; Honolulu, Hawaii; April 3, 2003 (Abstract).
Sandercock P. Should I start all my ischaemic stroke and TIA patients on a statin, an ACE inhibitor, a diuretic, and aspirin today? J Neurol Neurosurg Psychiatry. 2003; 74: 1461–1464.
Ovbiagele B, Saver JL, Fredieu A, Suzuki S, McNair N, Dandekar A, Razinia T, Kidwell CS. PROTECT: a coordinated stroke treatment program to prevent recurrent thromboembolic events. Neurology. 2004; 63: 1217–1222.
Sacco RL, Adams R, Albers G, Benavente O, Furie K, Goldstein LB, Gorelick P, Halperin J, Harbaugh R, Johnston SC, Katzan I, Kelly-Hayes M, Kenton EJ, Marks M, Schwamm LH, Tomsick T; American Heart Association/American Stroke Association Council on Stroke; Council on Cardiovascular Radiology and Intervention; American Academy of Neurology. Guidelines for prevention of stroke in patients with ischemic stroke or transient ischemic attack: a statement for healthcare professionals from the American Heart Association/American Stroke Association Council on Stroke: co-sponsored by the Council on Cardiovascular Radiology and Intervention: the American Academy of Neurology affirms the value of this guideline. Stroke. 2006; 37: 577–617.
Ovbiagele B, Saver JL, Fredieu A, Suzuki S, Selco S, Rajajee V, McNair N, Razinia T, Kidwell CS. In-hospital initiation of secondary stroke prevention therapies yields high rates of adherence at follow-up. Stroke. 2004; 35: 2879–2883.
Quilliam BJ, Lapane KL. Clinical correlates and drug treatment of residents with stroke in long-term care. Stroke. 2001; 32: 1385–1393.
Ovbiagele B, Saver JL, Bang H, Chambless LE, Nassief A, Minuk J, Toole JF, Crousse JR; VISP Study Investigators. Statin treatment and adherence to national cholesterol guidelines after ischemic stroke. Neurology. 2006; 66: 1164–1170.
Adherence Evaluation After Ischemic Stroke Longitudinal (AVAIL) Study. Available at: www.strokeassociation.org/presenter.jhtml?identifier= 3039649. Accessed July 26, 2007.
Bushnell CD, Goldstein LB. Physician knowledge and practices in the evaluation of coagulopathies in stroke patients. Stroke. 2002; 33: 948–953.
Watson RL, Dowell SF, Jayaraman M, Keyserling H, Kolczak M, Schwartz B. Antimicrobial use for pediatric upper respiratory infections: reported practice, actual practice, and parent beliefs. Pediatrics. 1999; 104: 1251–1257.