JCPSLP Vol 21 No 3 2019

Methods To provide an example of current practice and identify contextual gaps in service delivery, a retrospective cohort medical record review was conducted to identify key information. Ethical clearance was obtained from the appropriate human research ethics committee (HREC/17/ QPAH/722). Patients with dysphagia were identified retrospectively, with documented evidence of a complex dysphagia decision (defined as “risk” or “comfort” feeding) within the database resulting in a complete medical record review. The databases were reviewed within two medium- sized SLP departments for dates between January 2012 and December 2017, one being a private hospital with approximately 600 beds, the other a secondary public hospital with 435 beds, both in southeast Queensland, Australia. The full electronic- or paper-based medical record for the included individuals (presence of complex feeding decision) was then reviewed. Chart auditing was completed by key investigators using self-designed screening tools to identify the following outcomes: (a) patient demographic information and reason for admission; (b) severity of dysphagia and SLP recommendations based on the SLP’s documented impressions from their clinical swallow examination; (c) information on hospital admission including length of stay, discharge destination, adverse events/complications; and (d) information regarding complex feeding decision-making processes. Whenever possible, information regarding complex feeding decisions was recorded using a binary present/ absent method within the database. The information included the documented naming convention reported by SLP and medical teams (the exact working of the naming convention was copied verbatim into the data base and decision was made regarding exact agreement of wording between SLP and medical officer), the individuals documented to be involved in the complex feeding decision (i.e., patient, medical officer, SLP, interdisciplinary team), and documented education and/or interdisciplinary involvement (for example, documented evidence of a interdisciplinary case conference, provision of education materials, discussions regarding alternative feeding options). The entire medical record (including all discussions regarding complex feeding decisions) was reviewed for the patient’s admission. An exploratory approach to analysis was used in this discussion piece. Descriptive statistics included calculation of frequencies with percentages, means and standard deviations using IBM SPSS (IBM Corporation, 2016). Results Demographic information A total of 82 patient records were eligible for inclusion. The mean age was 78.75 (range 37–97, SD 12.16), with 46.3% (n = 38) being female and 53.7% (n = 44) being male. Reason for admission was most commonly due to multiple medical factors defined as a “complex” admission (26.8%, n = 22), followed by admissions for pneumonia (22%, n = 18). The majority of patients were admitted to hospital from their own home (54.9%, n = 45), followed by residential aged-care facilities (RACF) 39% (n = 32). The average length of stay for the current population was 13.10 days, (range 0–60, SD 11.55). The average number of SLP reviews was 5.11 (range 1–35, SD 5.34). Severity of dysphagia Prior to admission 63 patients had premorbid diet information available and 62 patients had premorbid fluid

consistency information available. The majority of patients were on a full or soft diet and thin fluids on admission to hospital; following the initial SLP review, a total of 51.2% were recommended to have nil diet orally and 46.3% were recommended to have nil fluids orally. Further information regarding diet and fluids on admission and following initial clinical assessment is shown in Table 1. At the time of initial SLP assessment, 41.5% (n = 34) of patients did not have a dysphagia severity diagnosis completed or were unable to be assessed, while 22% (n = 18) were assessed to have severe dysphagia, and 7.3% (n = 6) had moderate dysphagia. Only 8.5% (n = 7) of patients had an instrumental swallow assessment conducted. Complications such as pneumonia during admission (51.2%, n = 42) and malnutrition (29.3%, n = 24) were also identified.

Table 1. Diet and fluids on admission and following initial SLP review

On admission n (%)

Following initial SLP review n (% )

Diet

Full diet

20 (24.4%)

0 (0%)

Soft diet

17 (20.7%)

6 (7.3%)

Soft/minced diet

1 (1.2%)

0 (0%)

Minced-moist diet

5 (6.1%)

6 (7.3%)

Puree diet

16 (19.5)

19 (23.2%)

Nil diet orally

4 (4.9%)

42 (51.2%)

Fluid

Thin fluids

40 (48.8%)

9 (11%)

Mildly thick fluids

10 (12.2%)

14 (17.1%)

Tanya Hirst (top) and Lucy Lyons

Moderately thick fluids Extremely thick fluids

6 (7.3%)

7 (8.5%)

3 (3.7%)

5 (6.1%)

Nil fluids orally

3 (3.7%)

38 (46.3%)

** NB: n = 19 missing values for diet at time of admission, n = 20 missing values for fluids on admission, n = 9 missing values for fluids and diet post SLP review Complex feeding The majority of complex feeding decisions were documented as “risk feeding” (59.8%, n = 49) by the SLP, followed by “nil clear” documentation (19.5%, n = 16), and “other” risk feeding categories (12.2%, n = 10). Only 8.5% (n = 7) of complex feeding decisions were documented as “comfort feeding”. Only fair agreement was noted between the SLP’s documentation and the medical officer’s documentation with regards to complex feeding categories when comparing exact agreement between documented naming convention (kappa value 0.039, p = 0.189). Decisions regarding the commencement of risk/comfort feeding were made most commonly by the medical team, as shown in Figure 1. However, many of the complex feeding decisions were made without a documented discussion with the medical team (43.9%, n = 36), without a multidisciplinary case conference (82.9%, n = 68), without appropriate discussions regarding feeding decisions/options/alternatives (63.4%, n = 52), and without relevant patient/family education (59.8%, n = 49).

155

JCPSLP Volume 21, Number 3 2019

www.speechpathologyaustralia.org.au