Using rapid-cycle quality improvement methodology to reduce feeding tubes in patients with advanced dementia: before and after study

Context

At Lenox Hill Hospital, a 652 bed acute care facility in New York City, a multidisciplinary group of professionals concerned about improving medical care and quality of life for patients with dementia undertook a quality improvement project to address the issue of feeding tube placement in these patients. The project was conducted under the auspices of the New York Palliative Care Quality Improvement Collaborative (PCQuIC), a multisite initiative sponsored by the United Hospital Fund, RAND Corporation, and the Washington Home Center for Palliative Care Studies, with the aim of improving palliative care services in the 23 participating healthcare institutions.13 Faculty at the collaborative supplied training in the “plan-do-studyact” model for improvement14 15 and mentoring for this project over a 10 month period from September 2002 until June 2003. The project's core team included a geriatrician, a speech pathologist, a nurse, and an administrator. The hospital's vice president for medical affairs served as senior leader, acting as an advocate for the project.

Outline of the problem

Despite lack of evidence that feeding tube placement benefits patients with late stage dementia, despite medical risks and compromised quality of life, and sometimes contrary to the patient's advance directive, placement of feeding tubes in this population continues. We aimed to identify the extent and nature of this problem in our hospital and to intervene to change medical practice so that patients with dementia would receive medically appropriate treatment, consistent with their wishes. A retrospective chart review showed that within a six month period most feeding tubes were placed in patients with dementia. We found minimal documentation of patients' wishes regarding artificial nutrition and hydration, and few instances of formal assessment of the patients' capacity for decision making. Reversible causes for failure to eat were rarely addressed. The doctor's rationale for tube placement was documented infrequently. In cases of documented “failure to thrive,” no calorie count or other assessment had been made to determine actual food intake or to assess whether it was consistent with the nutritional needs of a patient with late stage dementia. A small but important number of patients received feeding tubes despite their explicit advance directives stating the wish to forgo artificial nutrition and hydration if they were irreversibly ill and unable to make their own decisions.

Key measures for improvement

We established three aims to measure the success of our interventions:

• Reduce by 50% the number of feeding tubes placed in patients with dementia

• Reduce to zero the number of feeding tubes placed in patients with dementia who are capable of taking food by mouth

• Reduce to zero the number of feeding tubes placed in patients with dementia who have advance directives stating the wish to forgo artificial nutrition and hydration.

We hoped to achieve these aims within nine months of initiating quality improvement interventions.

Methods used to identify problems

Names of all inpatients receiving percutaneous endoscopic gastrostomy or jejunostomy in the hospital from March to September 2002 were obtained from the hospital's computer database. The core team conducted an in-depth review of the 58 available charts of the 71 patients who received tubes. Data collected included patients' characteristics, medical diagnoses, presence of dementia (either reported as a diagnosis or inferred from notes in the chart), advance directives or other evidence of the patient's wishes regarding artificial feeding, doctor's rationale for tube placement, and a detailed chronology of the patient's hospital course. We conducted a second chart review, after initiation of quality improvement interventions, on all patients who received feeding tubes for the six month period March-September 2003.

Strategy for change

Beginning in January 2003, the team formed a palliative care consulting team and instituted educational programmes.

The palliative care consulting team was led by the director of geriatrics. Working with medical residents on geriatrics rotation and the existing ethics and pain consult services, the team provided support and guidance for attending physicians (figure). A letter from the hospital's vice president for medical affairs requested attending physicians to call a member of the team whenever a feeding tube was being considered for any patient, demented or otherwise. Primary care physicians, gastroenterologists, nutritionists, nurse managers, and case managers attended in-service training sessions that explained the project and enlisted participation (boxes 1 and 2) The project was publicised at the quarterly medical staff conference and at grand rounds.

Role of palliative care consulting team

• Download figure
• Open in new tab
• Download powerpoint

Four educational programmes were instituted. Firstly, a nationally recognised geriatrician with expertise in feeding issues in dementia gave a medical grand rounds presentation addressing the issue of feeding tube placement. The other three programmes involved the 24 medical residents in postgraduate year 2: they have a rotation of 12 sessions of pain management consultation; participate in presenting modules of the EPEC (Education for Physicians in End of Life Care) curriculum to the remainder of the medical house staff (attendance is mandatory); and spend one month on geriatrics rotation, which includes answering all palliative care consulting requests.

Effects of change

The table shows the preliminary results. The total number of tubes inserted and the number of tubes placed in patients with dementia both decreased significantly, exceeding our first aim, but the proportion of tubes placed in patients with an advance directive refusing artificial nutrition was not significantly reduced. This finding highlights a resistance by some doctors and surrogate decision makers to withholding treatment even when the patient's wishes to withhold treatment were clear. Retrospective data on ability to take food by mouth by patients in the 2002 group was not obtainable because few had received speech pathology or geriatrics evaluations to assess their ability to eat. Of the 2003 group of eight patients with dementia who received feeding tubes, two had been determined by the speech pathologist to be able to take some nutrition by mouth, four had been seen and determined to be unable to eat, and two had not received speech pathology assessments.

Lessons learnt

Use of rapid cycle quality improvement methodology13 14 to change medical practice in this case was successful on many levels. The intensive educational initiative, coupled with implementation of a palliative care consulting service, brought the issue of tube placement in patients with dementia into focus for the hospital community. Medical grand rounds and in-service training sessions were well attended. From January to September 2003, the consult team intervened on 12 feeding tube cases. It is unlikely, however, that the sharp reduction in tube placement is attributable only to the organised quality improvement interventions. During the study period, physicians, nurses, nutritionists, speech pathologists, case managers, social workers, and other clinical staff seemed to talk more, formally and informally, about feeding tubes and the wider issue of medically futile treatment, suggesting a “cultural shift” within the hospital. Staff from many disciplines took pride in the rapid, dramatic results of the project, and this increased morale and belief in the possibility of positive change within the institution. Doctors' response to the project was generally positive, although a few practitioners resented “interference” with their care of patients. More commonly, doctors were grateful for the help the consulting team offered in dealing with difficult cases. Education of house staff, in both didactic sessions and clinical rotations, was a key component in the success of the project.

We wondered if the sharp decline in feeding tube placement in patients with dementia in our hospital had been accompanied by an increase in outpatient tube placements in outpatient settings or at nearby facilities. We have joined a consortium of hospitals in New York City to address the issue on a regional basis.

It became clear from chart review and palliative care consultations that absence of clear goals of care often led to unnecessarily long, complicated stays in hospital, contradictions among recommendations made by various clinicians on a case, and care inconsistent with patients' wishes. The palliative care consultant became a facilitator of communication among medical staff and between doctor and patient or surrogate, to establish goals of care. We have started a new quality improvement activity for early establishment and documentation of goals of care.