Kathryn Head

1

THE DESIGN, IMPLEMENTATION AND EVALUATION OF A
VALID DYSPHAGIA SCREENING TOOL FOR USE BY
REGISTERED NURSES

• Kathryn Head
• Speech and Language Therapist Researcher

2
Identification of Problem-Literature Review

• Screening acute stroke patients for the
  presence/absence of dysphagia recommended within
  24 hours using a valid protocol
• No universally agreed screening criteria or tool
  exists exemplified by studies carried out
  worldwide
• Reasons cited limitations of validity due to
  differences in methodology, limited sample sizes.
  No single test is highly accurate for detecting
  dysphagia in patients and highly specific for
  ruling it out.

3

• Studies and reviews highlighted need to evaluate
  the minimum combination of predictive factors on
  their ability to detect dysphagia
• The potential of combining these factors in a
  dysphagia screening tool to evaluate if
  predictive for detection of dysphagia and for
  determining appropriateness of referral of acute
  stroke patients for bedside swallowing assessment
  by SLT not explored

4
Collection, Organisation and Interpretation of
Data

• Criteria reported in the literature as predictive
  for determining swallowing safety ascertained
• Inclusion/Exclusion criteria applied to
  literature review
• -sample sizes 100,
• -criteria/test compared against appropriate
  reference standard,
• -blinded, unbiased study design,
• -high-moderate to high sensitivity and
  specificity

5
Summary of Criteria with Reported Validity for
Determining Swallow Safety

• Patient alert/conscious to accept food and drink
• Upright posture
• Speed per swallowing a quantified volume of water
• Coughing during and following swallowing

6
Audit of Screening Practices in England and
Wales- February 2005

• Aim
• To evaluate the choice of tools used, range of
  nurse grades trained and the frequency of use of
  criteria reported in the literature as having
  moderate-high predictive validity for determining
  swallow safety
• Method
• An audit of 12 acute trusts in Wales and 40
  trusts falling within the 28 Strategic Health
  Authorities in England

7

• Lead SLTs for dysphagia contacted
• Asked whether screening takes place in Trust. If
  not why?
• Asked re range of nurse grades trained, if
  possible try to quantify
• Asked to provide copy of tool so that frequency
  of use of predictive criteria could be calculated

8
Outcomes

• 96 response rate
• 11 Trusts did not screen-reasons cited include
  lack of resources, lack of compliance from
  nurses, lack of time
• 34 out of 38 Trusts trained newly qualified RGNs
  and upwards
• All Trusts included alertness as a criterion and
  most included upright posture. Only 4 Trusts used
  speed of swallowing and few specified coughing
  during and following swallowing

9
Interpretation of Data

• Wide variation in dysphagia screening practices
• SLTS not fully informed by current evidence in
  selection of screening tools/criteria
• Need for a National framework to inform selection
  of criteria and training of nurses
• Need for evaluation of minimal combination of
  predictive criteria on their ability to
  accurately determine appropriate referral of
  dysphagic patients to the SLT

10
Action based on dataSept 2006-July 2007

• A bedside dysphagia screening tool for nurses
  designed, incorporating the four screening
  variables identified as predictive for
  determining the presence or absence of dysphagia

11
(No Transcript)
12
Purpose of Multi-phased Research

• Purpose to identify and evaluate the concurrent
  and predictive validity of a minimum combination
  of evidence based dysphagia screening within a
  prototype dysphagia screening tool.

13
The Reliability of the SLTRs Bedside
Assessment of Swallowing

• Aim
• To determine that the SLTRs bedside assessment
  of swallowing is a reliable and appropriate
  reference standard against which to measure the
  validity of the prototype dysphagia screening
  tool
• Objective
• To measure the inter-rater reliability of the
  SLTR assessment of swallowing as compared to a
  similarly qualified SLT contemporys assessment
  for determining the presence or absence of
  dysphagia in a prospective sample of 30 acute
  medical patients referred to the SLT department.

14
Methods

• Bedside assessment of swallowing performed by two
  Clinical lead SLTs on a sample of 30 acute
  medical patients including acute stroke patients.
• SLTs blinded to one anothers assessments
• Specific criteria used within each SLTs
  assessment analysed after data collection to
  determine congruence or differences

15
Results

• Twenty-six patients out of 30 assessed received
  the same assessment ratings i.e. dysphagia
  present or dysphagia absent.
• The proportion of agreement for the two
  therapists dysphagia assessment was calculated
  as 87. Kappa was calculated as 0.71 suggesting
  substantial agreement.

16
Conclusion

• Results suggest an inherently high inter-rater
  reliability.
• Differences due to changes within the patients
  and a different emphasis placed on some clinical
  signs observed.
• SLTRs bedside assessment was an appropriate
  reference standard against which to compare the
  performance and use of the screening tool

17
Inter-rater Reliability Study of Prototype
Dysphagia Screening Tool

• Aim
• To evaluate whether RGNs agree with the SLTR when
  both are using the research-screening tool. This
  phase was designed to elucidate potential
  problems with the nurses training programme or
  design of the tool.
• Sample
• SLT Researcher and two representative grade RGNS
  (one novice RGN and one experienced RGN).
• 40 acute stroke patients

18
Methods

• Blinded study carried out on an acute stroke
  ward.
• Novice RGN undertook screening using the
  prototype dysphagia for twenty acute stroke
  patients followed by the SLTR performing
  screening again using the prototype dysphagia
  screen on the same 20 patients.
• Procedure repeated with Experience RGN and SLTR
  with another convenience sample of 20 acute
  stroke patients.
• Once all data collected, analysed to determine
  the proportion of agreement

19
Results

• 35 patients out a total sample of 40 screened,
  received the same ratings i.e. dysphagia present
  or dysphagia absent and patient appropriate for
  referral to SLT patient/not appropriate for
  referral.
• SLTR vs Novice RGN agreed on 17 out of 20
  patients screened proportion of agreement
  calculated as 85. Kappa 0.71 for dysphagia
  detection, Kappa 0.79 (plt0.001) for determining
  appropriateness of referral suggesting high
  agreement
• SLTR vs Experienced RGNagreed on 18 out of 20
  patients proportion of agreement as 90. Kappa
  0.77 plt0.001 for detection of dysphagia and
  Kappa for determining the appropriateness of
  referral 0.87 plt0.001

20
Conclusion

• The results of this small study suggested that
  the research dysphagia-screening tool in the
  hands of the representative grade nurses was
  highly reliable for determining signs of
  dysphagia and for determining patients
  appropriate for referral to SLT for full
  assessment of swallowing.
• It is argued that these findings supported
  proceeding to the next and final phase of the
  research programme, to measure the validity of
  the screening tool in the hands of nurses
  compared to a full bedside assessment of
  swallowing performed by the SLTR.

21
An Evaluation of the Validity of the Prototype
Dysphagia Screening Tool

• Aim
• To test the hypothesis that the research
  dysphagia screening tool is a valid tool for
  determining acute stroke patients appropriate for
  assessment of dysphagia by the SLT.
• Objective
• To measure the concurrent and predictive validity
  of the research dysphagia screening tool when
  used by nurses compared to a full bedside
  assessment of swallowing performed by an expert
  SLT (the SLTR) in determining the presence or
  absence of signs of dysphagia and judging whether
  the patients are appropriate for referral for
  full bedside assessment of swallowing.

22
Study Design

• Comparison of the detection of dysphagia and
  appropriateness for referral of the patient for a
  SLT swallowing assessment as determined by the
  prototype dysphagia screening tool performed by
  two RGNs compared against the full bedside
  assessment of swallowing by the SLTR
• Population and Sample
• Two RGNs (different from the previous phase) both
  worked on acute medical wards with stroke
  patients Novice RGN a basic grade RGN with
  18 months post qualification working experience,
  Experienced RGN RGN with 15 years working
  experience
• SLTR with 12 years postgraduate dysphagia
  experience and postgraduate training in dysphagia
  to masters level equivalent.

23
Methods

• Blinded study design
• Within a working day, each nurse independently
  screened the same acute stroke patients using the
  prototype dysphagia screen and recorded their
  screening outcomes for detection of dysphagia and
  appropriateness of referral for full bedside
  assessment.
• Followed by SLTR full bedside assessment on the
  same acute stroke patients. Outcomes again
  independently recorded.

24
Results-Detection of Signs of Dysphagia

• SLTR Vs Experienced RGN Correlated 89 for
  determining dysphagia presence and absence.
  Sensitivity 0.88, specificity 0.87 plt0.001
• SLTR Vs Novice RGN Correlated 91 for
  determining dysphagia presence and absence.
  Sensitivity 0.88, specificity 0.95 plt0.001

25
Results-Determining Appropriateness of Referral
for Full Swallow Assessment

• SLTR Vs Experienced RGN Correlated 89
• Sensitivity 88, Specificity 90
• SLTR Vs Novice RGN Correlated 90
• Sensitivity 85, Specificity 95

26
Summary

• The study has identified a valid combination of
  dysphagia screening criteria within a minimally
  invasive dysphagia-screening tool, for use by
  RGNs.
• The screening tool has high accuracy for
  determining the presence and absence of dysphagia
  and for identifying patients appropriate for full
  assessment of swallowing by the SLT.
• First study to evaluate concurrent and predictive
  validity of evidence based screening tool for use
  by RGNs