North Glasgow

1
North Glasgow what were the reasons for
discrepancies between SMR01 and audit data

• Dr M Taylor, Dr C McAlpine, Department of
  Geriatric Medicine, Stobhill Hospital, Glasgow

2
Background

• ISD (Information Services Division) collects data
  on the number of patients treated for stroke and
  the procedures they receive
• This together with data on deaths allows ISD to
  monitor mortality, incidence and treatment of
  stroke
• SSCAS uses ISD data as a comparison

3
Background

• ISD and SSCAS class cerebrovascular disease as
  follows (according to ICD-10)
• CVD I60-9, G45
• Stroke I61-4
• SAH I60
• TIA G45

4
Background

• Inconsistency in coding can result in
  incomparability of data
• Can be caused by misunderstanding of rules by
  individual coders
• Different methods and approaches to the
  collection of information at source
• ISD has specific standards about quality of data

5
Background

• Looking at the SSCAS figures for last year it was
  thought that there were some discrepancies in the
  figures for Stobhill Hospital ( both the data
  collected and the ISD figures SSCAS had )
• This was thought to be the case even when
  accounting for strokes discharged under
  non-stroke physicians

6
Background

• Stobhill Hospital has an incremental discharge
  letter system (IDLS) that can code on immediate
  discharge
• The IDLS is reviewed by the consultant and a
  final discharge letter is then formed
• The IDLS system does not allow for adjustment of
  diagnosis
• On reviewing the IDLS document many strokes were
  classified as G81.- (hemiplegia) rather than
  Stroke I61-4.-

7
Aims

• To look at the difference in IDLS coding and
  actual coding could this be a source of error?
• To try and account for any discrepancies in the
  recorded number of strokes according to ISD /
  SSCAS and the perceived number of strokes

8
Methods

• Data was gathered from discharge summaries over a
  6 month period (Oct 2004 March 2005) from the
  stroke service team
• IDLS coding was recorded where applicable
• I coded the discharge summaries according to ICD
  10
• Actual Trust ICD 10 coding was then recorded in
  those deemed to have a diagnosis of stroke

9
Results

• 130 discharge summaries were evaluated in total
  from the stroke service

10
Results - Demographics - Sex
11
Results - Demographics - Age
12
Results Stroke type
13
Results

• IDLS was completed on 72 out 130 (55) of
  discharges
• 38 of those 72 (53) strokes were not coded as
  stroke on IDLS
• Of these 38, 26 (68) were then coded by coding
  dept as stroke
• 1 stroke on IDLS was coded as not stroke by
  coding department

14
Results ICD 10 Comparison
15
Results

• 127 patients were classified as stroke disease
  when reviewing the discharge summaries compared
  to 45 according to SSCAS and 97 according to ISD
  (for a longer time period)
• 13 patients who should have been coded as stroke
  according to discharge letter were not

16
Actual Diagnosis Final ICD 10 Coding Diagnosis
R PACI (I63.9) Headache, malaise, other unspec neurology
Stroke (I64.X) Hemiplegia, unspecified
Stroke (I64.X) TIA, unspecified
Stroke (I64.X) Congestive heart failure
Stroke (I64.X) Hemiplegia, unspecified
Stroke (I64.X) TIA, unspecified
Stroke (I64.X) Occlusion and stenosis of MCA
Stroke (I64.X) Dysarthria and anarthria, LVF
R LACI (I63.9) Not coded
Stroke (I64.X) Other unspecified neurological condition
Stroke (I64.X) Ess hypertension, constipation, mild cogn imp
L PACS (I63.9) Acute psychosis, AF, ess. hypertension
L PACS (I63.9) Syncope and collapse
17
Conclusions

• There are some areas which are of concern namely
• Large discrepancy between SSCAS stroke figures
  for Stobhill and our figures
• Strokes being coded as hemiplegia (G81.-),
  dysarthria (R47.1) and collapse (R55.X) which are
  symptom related codes as opposed to disease
  related
• G45.- (TIA) used instead of I61-4 (Stroke)
• Strokes not being classed as any form of
  cerebrovascular disease at all

18
What have I gained from this?

• Coding is an important issue (not boring as I
  previously thought!)
• Implications for future healthcare resources and
  planning if data incorrect
• There are several potential sources of error in
  the data collection / coding process

19
Further Plans

• To review IDLS coding at Stobhill
• To suggest that hemiplegia (G81) is not selected
  as stroke on IDLS
• To look further into where the SSCAS team gain
  their figures from for Stobhill
• To look at other stroke departments throughout
  the Trust

20
Any Questions?