Audit of stroke care at the Institute of Neurology , National Hospital of Sri Lanka , 1994-1997

Institute of Neurology, National Hospital of Sri Lanka,Colombo. 2 Faculty of Medical Sciences, University of Sri Jayawardenapura, Nugegoda, Sri Lanka. (Received 5 January 2002; revised version accepted 31 May 2002. Corresponding author UKR, telephones +94 1 674681, +94 1 91790, email: udayaran@yahoo.com). September 1997) were scrutinised, and records of all patients with stroke (n=263) were identified for retrospective analysis. No sampling was performed. Case selection was based on the clinical diagnosis of stroke based on WHO criteria (2), and recorded confirmation by computed tomography (CT) where available. Cases with an initial clinical diagnosis of stroke, all subsequently refuted by CT were excluded. Patients with transient ischaemic attacks (TIAs) and patients readmitted for further investigation or management of complications of a previous stroke were also excluded.


Introduction
Stroke is a much neglected illness.In spite of the enormous impact of stroke on patients, care givers and the society, stroke care remains poor throughout the world, even in developed countries (1).Stroke care in Sri Lanka has not been previously evaluated.We did an audit of stroke care at the Institute of Neurology of the National Hospital of Sri Lanka (NHSL), Colombo.

Audit package
An audit package designed by the Royal College of Physicians of London (RCP) and the United Kingdom Stroke Audit Group (3) was used to evaluate the process of stroke care.The package consisted of stroke audit forms for the data entry, and software for data analysis.The RCP stroke audit package has been previously used in many centres (4,5).

Data collection
Data from the case records were recorded in RCP stroke audit forms and trasferred into a computer database.The audit form contained 60 audit items, categorised under 13 subheadings assessing different aspects of documentation of stroke care (Table ).
There were 3 possible responses to each audit item; 'Yes' -if the information has been recorded, or 'No' -if the information has not been recorded, or 'No, but' -if the information has not been recorded, but this was considered acceptable in the given circumstance.For example, not assessing visual fields or speech in an unconscious patient would be recorded as a 'No, but' response.The relevant information had to be recorded within a specified time frame to be considered acceptable.

Data analysis
The result for each audit item was expressed as a percentage value (OK%), which was an indication of the quality of documentation of care.For the purpose of this study, we graded quality of care according to the OK% as follows: very good (81 to 100%), good (61 to 80%), average (41 to 60%), poor (21 to 40%), and very poor (0 to 20%).

Results
'Very good' care as defined for the study was achieved only in 11 (18.3%) of the audit items.These included certain neurological assessments (such as recording of symptom evolution, blood pressure, conscious level, limb power, funduscopy), performing baseline investigations (such as blood count, urea and electrolytes, blood glucose), and consultant review.'Good' care was noted for only 9 (15%) of the audit items.Carrying out routine investigations such as erythrocyte sedimentation rate (ESR) and ECG, computed tomography (CT) scanning, investigating for rare causes of stroke and starting aspirin for secondary prophylaxis were among these.
Care was 'average' for 5 (8.3%), 'poor' for 9 (15%) and 'very poor' for 26 (43.3%) of the items.Basic clinical data such as drug history, previous strokes or TlAs, peripheral vascular disease, smoking and alcohol habits, and cardiac rate and rhythm were poorly documented.Assessment of aspects important in rehabilitation such as pre-stroke function, communication, trunk control or gait, swallowing, visuospatial function and visual fields was extremely unsatisfactory.Formulation of a clear and complete diagnosis, and immediate management of hydration and urinary incontinence were deficient.Many aspects related to rehabilitation planning, discharge planning and secondary prevention were very poor (Table ).

Discussion
We used a previously validated audit instrument, the RCP audit package, for this study.The RCP stroke audit was simply an audit of good clinical practice.The majority of the items audited (51 out of 60) are basic clinical variables that can be easily assessed by the bedside, and the routine investigations are readily available in most hospitals.Facilities for CT scanning, non-invasive carotid imaging and investigation for unusual causes of stroke (echocardiography, haematological screening etc.) were available at the NHSL during the period of study.The RCP audit package was therefore applicable in our setting.
Audit is a powerful tool for evaluating patient care in clinical practice, enabling identification of relative strengths and weaknesses.This study indicated serious deficiencies in stroke care at the Institute of Neurology.Rehabilitation oriented clinical assessments, rehabilitation planning and discharge planning were extremely poor.This may have been partly due to the paucity of comprehensive rehabilitation services in acute care settings, the complete absence of a multidisciplinary approach to patient care, and the lack of emphasis on rehabilitation in undergraduate and postgraduate medical education.
The purpose of audit is to identify deficiencies of care, and to introduce remedial interventions.As a result of this audit, several changes were made in the structure and process of stroke care at the Institute of Neurology.These included the introduction of a structured clerking format to improve documentation, including rehabilitation oriented neurological and functional status assessments.A widely used functional assessment scale, the Barthel Index of activities of daily living (6), was included in the clerking format.A stroke unit was established at the Institute.A multidisciplinary stroke team was set up, with specialist inputs in speech therapy, physiotherapy, occupational therapy and social services support.

Conclusion
The RCP stroke audit package was successfully used at the Institute of Neurology, and stroke care was found to be suboptimal in many aspects.The audit has served to bring about major changes, which we believe will lead to significant improvements in stroke care.

2 (
all patients admitted to the Institute of Neurology over a period of 3 years (October 1994 to Audit of stroke care at the Institute of Neurology, National Hospital of Sri Lanka, 1994-1997 U K Ranawaka 1 , M A H Fernando 1 , T G S M Gunasekera 1 , B S N Senanayake 1 , D R R Gunawardena 1 , J C Wijesekera 1 , and D J S Fernando Index words: Retrospective case record evaluation, deficiencies, interventions) September 1997) were scrutinised, and records of all patients with stroke (n=263) were identified for retrospective analysis.No sampling was performed.Case selection was based on the clinical diagnosis of stroke based on WHO criteria (2), and recorded confirmation by computed tomography (CT) where available.Cases with an initial clinical diagnosis of stroke, all subsequently refuted by CT were excluded.Patients with transient ischaemic attacks (TIAs) and patients readmitted for further investigation or management of complications of a previous stroke were also excluded.