Abstract
Abstract
The purpose of this translation of research into practice (TRIP) project was to to determine the impact of a multidisciplinary education-consultation intervention to reduce percutaneous endoscopic gastrostomy (PEG) tube placement in patients with terminal-stage dementia at a single urban hospital in a city characterized by numerous health care transitions. We attempted a “just-in-time” approach to educate busy clinicians through explicit recommendations offered during routine and requested consultation. The project results showed that the intervention had a modest positive clinical impact.
Introduction
Major adverse effects associated with PEG tube placement across all patients include aspiration, peritonitis, hemorrhage, tube migration, gastrocolocutaneous fistula, wound infection, necrotizing fasciitis, and inadvertent removal of the tube. Minor complications include leakage and blockage. 3 In spite of the growing body of evidence that patient benefit is unlikely and does not exceed the risk of minor or major complications, PEG tube placements remain relatively common.
Teno and colleagues reported a 53.9/1000 incidence of PEG tube placement among 97,241 nursing home residents in the United States with 48.3% of the cohort dying within a year after PEG tube insertion. Additionally, a strongly significant relationship between the number of health care transitions and PEG tube placement was found. In other words, nursing home residents who experienced more transfers to or from the hospital, inpatient rehabilitation, inpatient hospice, and home care with or without hospice were more likely to have a PEG tube placed than patients with few transfers. 4 Patients experiencing numerous transfers are likely to encounter many providers across numerous settings who have varying knowledge or attitudes regarding PEG tube placement in this patient cohort.
Nursing home residents with advanced dementia were followed for up to one year to see whether a PEG tube was placed and for up to one year postinsertion to examine survival and health care use. Investigators found a 64% one-year postinsertion mortality with a median survival after PEG placement of 56 days. Additionally, 19% required tube repositioning within a median of 4.8 months after insertion. Lastly, 68% of tubes were placed in an acute care hospital making the hospital providers the primary decision makers with patient's family or surrogates. 5
Decisions to start PEG tube feedings were associated with internal medicine specialty, family preference, and fear of liability in a survey of primary care physicians in Hawaii. 6 These results suggest that the evidence about morbidity, mortality, and nonbenefit are either not well known or carry a smaller weight in physician decision making than other factors.
The decision-making process is complicated, and involves additional issues such as patient's advance directives, legal concerns or mandates, financial incentives to place rather than withhold tubes, surrogate emotional factors, cultural background, religious beliefs, and the adequacy of clinician communication. 7 For this translation of research into practice (TRIP) investigation we focused on only one factor, that is, the knowledge of the clinician about the evidence of nonbenefit in this patient population. The clinician brings credibility and science to family meetings when treatment options are being considered. The clinician who merely says “your mother can't swallow, do you want us to put in a feeding tube?” may be lacking knowledge about the evidence. The intuitive family response will be to say yes.
It may take as long as one or two decades to translate original research into practice. 8 Thus, the purpose of this TRIP project was to determine the impact of a multidisciplinary education-consultation intervention to reduce PEG tube placement in patients with terminal-stage dementia at a single urban hospital in a city characterized by numerous health care transitions.
Method
After obtaining Human Investigation Committee approval, a quasi-experimental comparison was conducted from retrospective chart review. Waiver of consent was approved because patient identifiers were used but not disclosed and results are reported in the aggregate. Records of inpatients admitted between January 1, 2008 and February 28, 2009 with a history of dementia were reviewed; the multidisciplinary intervention began in August 2008. Patients admitted from January 1, 2008 through February 28, 2009 who did not receive the intervention composed a historical control group. Patients admitted between September 1, 2008 and February 28, 2009 who received the intervention constituted the intervention group.
The impetus for this TRIP began following a nursing grand rounds lecture in January 2008 summarizing the evidence of nonbenefit of enteral feedings in patients with advanced dementia. There were no physician attendees at this session, thus no prior exposure from the study team about the evidence. Of attendees, an interested multidisciplinary team was formed comprised of palliative care nurse practitioners, the speech language pathologist (SLP), and the registered dietitians to conduct this TRIP project and measure the frequency of PEG tube placement as the primary outcome. Nurse practitioners from the palliative care service frequently counsel patients' families about PEG tube placement during patient hospitalization; however, not all patients with advanced dementia are referred for palliative care consultation. Hence, the inclusion of the SLP and dietitians, who routinely see patients admitted with malnutrition, dehydration, and/or dysphagia, the common admitting diagnoses in advanced dementia, was expected to be useful.
The intervention was not aimed at the multiple and varied reasons that PEG tubes are placed in patients with advanced dementia. We simply educated the attending physician team about the nonbenefit of PEG tubes in this patient population during routine or requested consultation. Additionally, we explicitly identified the patient as having “advanced” or “terminal” stage dementia to provide the clinical context and bring awareness of the terminality of advanced dementia.
Sample
Patient medical records were reviewed by a registered nurse research assistant (RA). Records were selected for review with the following inclusion criteria: Patients were admitted to the hospital between January 1, 2008 and February 28, 2009 and had a history of some form of dementia; we searched records by Diagnosis Related Group codes (see Table 1). Patient records were examined for documentation of advanced dementia and patients were included in subsequent analyses if they had no PEG tube in place at hospital admission. (See Fig. 1 for sample distribution and study inclusion). During the study interval 71 patients with advanced dementia constituted the historical control group, whereas 10 patients received the intervention.
Sample distribution.
NOS, not otherwise specified.
Intervention
The educational intervention in this project consisted of a “just-in-time” identification of the advanced-stage dementia and an explicit recommendation to withhold PEG tube placement to the referring physician. The registered dieticians routinely provide consultation to any patient who is admitted with nourishment problems. The SLP and palliative care providers provide consultation at the request of the attending service. After assessing the patient for clinical evidence of advanced dementia, all project team members provided advice about PEG tube nonbenefit to the referring physician when consulted for swallow evaluation, nutritional assessment, or palliative care. We mutually agreed to write discipline-specific impressions and recommendations; the SLP and dietitians recommended consultation with palliative care. (See Table 2 for sample consultant documentation.)
pt, patient.
During the TRIP interval there were no significant local or national feeding tube cases receiving media attention, no other grand rounds sessions (medical or nursing) focused on PEG tube and dementia evidence, and no enhancements to the electronic medical record to flag the provider about PEG tube evidence; thus, we could not identify any external or historical threats to our intervention.
Outcome measures
Demographic data, including, age, gender, and race were recorded. Additionally, the patient's admitting diagnosis, estimated dementia stage, hospital length of stay, disposition, and hospice referral status were recorded. Consultations, if any, with a dietitian, SLP, geriatrician, gastroenterologist (GI), surgeon, or palliative care provider were noted along with their recommendations regarding PEG tube placement. We further identified whether a PEG tube was placed prior to patient discharge.
Dementia staging was the most labor-intensive aspect of the chart review because a standardized assessment with an instrument such as the FAST (Functional Assessment Staging Tool) 9 is not routinely done on all demented patients. Evidence from the history and physical, nursing admission assessment, and other clinical documentation enabled categorizing the patient into an estimated FAST score <7 or ≥7. Estimated FAST scores ≥7 signified advanced dementia for this investigation. For example, if the patient documentation indicated that the patient was diapered; had pressure ulcers, contracted lower extremities, and other physical evidence of being nonambulatory; was incontinent; and was confused we categorized the patient as having advanced dementia. The research assistant (RA) did the initial review, and the principal investigator (PI) conducted a subsequent review of 10% of all cases and of all cases in which the RA was uncertain to establish reliability in the estimation of dementia stage. Patients were excluded if the documentation was ambiguous to both the RA and PI.
Hospital length of stay was counted using a calendar day with day one as the day of admission and last day as the day of death or discharge. Disposition was categorized as died, home, or long-term care (LTC). Consultants' recommendations regarding PEG tube were categorized as none, place PEG tube, and do not place PEG tube.
Results
The Statistical Program for Social Sciences (SPSS, v. 18; SPSS Inc., Chicago, IL) was used for data entry and analysis. During the project interval we identified 81 patients with advanced dementia who did not have a PEG tube prior to admission. The most common admitting diagnosis was dehydration in 27% of patients; other expected diagnoses such as pneumonia, urinary tract infection, and septic shock were seen with lesser frequency.
Table 3 summarizes the sample characteristics. The sample was comprised of slightly more women than men and no significant differences in age, race, or gender between the control and experimental groups were found. Additionally, there were more African American patients than other racial groups, which reflects the urban demographics of the hospital. More patients were referred to hospice at discharge from the experimental group, although the difference between groups was not statistically significant.
SD, standard deviation; PEG, percutaneous endoscopic gastrostomy; FAST, Functional Assessment Staging Tool; LTC, long-term care.
Table 4 illustrates the frequency of consultations and consultants' recommendations regarding PEG tube placement. Although the intervention group is quite small, we achieved a significant increase in explicit recommendations to withhold PEG tube placement from all project team consultants and we saw no difference in recommendations made by GI, geriatrics, or surgery consultants. Of patients in the control group, 4 of 71 (5.6%) had a PEG tube placed prior to discharge, whereas no patients in the experimental group received a PEG tube, but the difference is not statistically significant (×2 = 0.59, p = 0.44). Notably, one patient received GI consultation with a recommendation to place a PEG tube; however, the patient also had a recommendation from the dietician to withhold PEG tube placement and subsequently the patient did not have a PEG tube placed. No significant gender or racial differences were found between those with a PEG tube placed and those without.
PEG, percutaneous endoscopic gastrostomy; GI, Gastroenterologist.
Discussion
In this project our goal was to make our clinical colleagues aware of the evidence about nonbenefit of PEG tube placement in advanced dementia. Although this evidence is commonly known and understood by palliative care providers and geriatricians, it may be less well understood by other clinical disciplines. We attempted a just-in-time approach to rapidly and succinctly educate busy clinicians through explicit recommendations offered during routine and requested consultation.
Our rate of PEG tube placement is consistent with other reports. 4 Buy-in from the registered dietitians and the SLP was integral to the success of the intervention by assuring that one of the possible consultants involved in this intervention would see the patient. Dietitian consultation is automatic in this hospital when the patient is receiving nutrition support, has an open or nonhealing wound, is at risk for pressure ulcers, has difficulty chewing or swallowing, or when the patient's intake appears insufficient or clinical course implies risk. Palliative care consultants or the SLP do not routinely see patients without a referral from the attending team. Concerns about scope of practice, accuracy in identifying advanced dementia, and phrasing recommendations for the various disciplinary recommendations were discussed and reconciled as a study team a priori.
We found no statistical significance between groups with regard to PEG tube placement, and this is certainly because of a very small sample, particularly in the intervention group. However, a clinical trend toward fewer PEG tube placements in this patient population can be predicted as we continue the intervention. The palliative care team is receiving requests for supportive documentation about the nonbenefit of PEG tubes that we have not received previously. Most of the physicians who admit patients to this hospital are academic faculty, and although the trainees rotate to other sites, the attending/faculty is relatively stable, which should enhance the enculturation of PEG tube evidence.
As others have reported, dementia is often not recognized as a terminal illness. 10 In this intervention the consultant explicitly identifies the terminal status of the patient. An unexpected finding was a significantly increased frequency of palliative care consultation in the experimental group, which may have resulted from the secondary recommendation of the registered dietitian or the SLP to consider consultation with palliative care. The greater frequency of palliative care consultation also may have influenced the trend toward an increased hospice referral at discharge compared with the control group.
This TRIP intervention has merit but is not without limitations. The intervention was conducted at a single, academic institution where there is a seasoned palliative care consultation service of 25 years duration and an enculturation of evidence-based practice. Other settings with a more heterogeneous physician staff and/or little or no academic or palliative care presence may not find the same results. Further, we relied on retrospective chart review for our data retrieval. The intervention itself was limited in that we only addressed the clinical evidence about PEG tube placement in dementia and we did not address other factors such as family preferences, religion, culture, financial gain, or physicians' fears about liability.
Summary
A multidisciplinary consultation-education intervention to translate PEG tube placement in advanced dementia evidence into practice has had a modest positive clinical impact. Additionally, the intervention was easily implemented with no inherent costs. Explicit just-in-time consultant documentation of the terminal stage of dementia and the nonbenefit of PEG tube placement and enteral feedings fostered the translation of evidence into practice.
Footnotes
Author Disclosure Statement
No competing financial interests exist.