Feeding tubes in advanced dementia

    Whether a person with advanced dementia who has trouble eating should have a feeding tube, or continue careful hand feeding.

    Families and others deciding for an adult with advanced or severe dementia who now eats or swallows poorly. Most of the evidence comes from US nursing home residents.

    SourcedModeledNo projected outcomes are generated on this page.

    Pending clinical review. Each statement is paraphrased from the named source, but no clinician has reviewed this page yet. Use it to prepare questions, not to choose a treatment.

    What this decision concerns

    For people with severe dementia, the balance of benefits and harms of tube feeding is not clear. Its use is falling in some parts of the world and rising in others.

    SourcedCochrane systematic review · 2021

    The American Geriatrics Society does not recommend feeding tubes in advanced dementia and offers careful hand feeding instead. It also calls tube feeding a medical therapy that the person's decision-maker can accept or decline, following advance directives, earlier wishes, or what the person would want.

    SourcedAmerican Geriatrics Society · 2014

    The options

    Place a feeding tube

    Nutrition and fluids through a tube into the stomach (gastrostomy) or through the nose (nasogastric).

    • Three of four studies of mixed or unspecified tube feeding found tube feeding associated with longer survival; the fourth, the largest, found no effect. Certainty is very low.

      Applies to: Non-randomized studies; 310 participants in the three, 1386 in the fourth.

      SourcedCochrane systematic review · 2021

    • Two studies gave very-low-certainty evidence that tube feeding improved nutritional measures.

      Applies to: Non-randomized studies of adults with severe dementia.

      SourcedCochrane systematic review · 2021

    • Four studies of gastrostomy (PEG) tubes found no evidence of an effect on survival time (low certainty).

      Applies to: 36,816 participants in four non-randomized studies.

      SourcedCochrane systematic review · 2021

    • Survival did not differ between residents who had a gastrostomy tube placed and those who did not, and placing it sooner after eating problems began did not change survival.

      Applies to: 36,492 residents, 1,957 of them tube-fed. US nursing home residents with advanced cognitive impairment from dementia and new problems eating, 1999 to 2007.

      SourcedUS national nursing home cohort study · 2012

    • Residents who had a gastrostomy tube placed in hospital were 2.27 times as likely to develop a new pressure ulcer, and existing ulcers were less likely to heal.

      Applies to: Propensity-matched US nursing home residents with advanced cognitive impairment who recently needed help eating; 1124 with a tube placed for new ulcers, 461 for healing.

      SourcedUS national nursing home cohort study · 2012

    • On balance, the studies suggested a higher risk of pneumonia with tube feeding, though results were inconsistent.

      Applies to: Five non-randomized studies reporting harms.

      SourcedCochrane systematic review · 2021

    • Tube feeding is associated with agitation, more use of physical and chemical restraints, and health care visits for tube complications.

      Applies to: Older adults with advanced dementia, as summarized in the position statement.

      SourcedAmerican Geriatrics Society · 2014

    • In one study, more carers in the tube feeding group reported a very heavy burden. The same study found no difference in pain or comfort.

      Applies to: One non-randomized study using mixed tube feeding methods.

      SourcedCochrane systematic review · 2021

    Continue careful hand feeding

    Food and drink by mouth, offered with help, at the person's pace, for comfort.

    • Hand feeding has been shown to be as good as tube feeding for death, aspiration pneumonia, functional status and comfort.

      Applies to: Older adults with advanced dementia, as summarized in the position statement.

      SourcedAmerican Geriatrics Society · 2014

    • Adjusting the eating environment and the way food is offered to the person should be part of usual care.

      Applies to: Older adults with advanced dementia.

      SourcedAmerican Geriatrics Society · 2014

    • In one large study, people not given a gastrostomy tube had a lower risk of pressure ulcers (moderate certainty).

      Applies to: 4421 participants, 1585 with a tube and 2836 without, in one non-randomized study.

      SourcedCochrane systematic review · 2021

    • In three of four studies of mixed or unspecified tube feeding, people who were not tube-fed survived for less time. Certainty is very low.

      Applies to: Non-randomized studies; 83 participants without tube feeding in the three.

      SourcedCochrane systematic review · 2021

    Set against what matters to you

    Modeled

    Read from what you saved on this device.

    What sources report about what matters to you

    Nothing yet. Mark what matters to you, and anything the sources say about it shows up here.

    Pulls in two directions

    Nothing so far pulls against anything else.

    Still unknown

    • Nothing is marked as a priority yet, so nothing here is connected to what matters to you. The options and their sources are still readable below.

      Not in the sources here.

    • No randomized trial has compared tube feeding with hand feeding in severe dementia. All 14 studies in the Cochrane review were observational, with high risk of bias from confounding.

      Cochrane systematic review, 2021 statement

    • None of the studies measured quality of life.

      Cochrane systematic review, 2021 statement

    • Most of these studies come from US nursing homes. Whether the findings hold in hospitals, at home or in other countries is not answered here.

      Not in the sources here.

    • Some faith and cultural traditions hold views about nutrition at the end of life. Those views are not part of this evidence record.

      Not in the sources here.

    Worth asking

    • What has the person said, in writing or out loud, about tube feeding or about care near the end of life?
    • What is causing the eating problem now, and could any of it change?
    • What would careful hand feeding look like here, and who would do it?
    • If a tube were placed, what would change in daily care, and how would we know if it was helping?
    • Whichever we choose, how will comfort be checked, and when will we talk about it again?
    Start with what matters

    How this measures up

    Checked against IPDAS 5.0: 10 met, 1 partly met, 3 not met

    Sets source-backed statements about named options against what the person said matters. Does not meet the development and evaluation standard a patient decision aid requires. Three further stages would be needed to call it a patient decision aid: essential-criteria appraisal, an independent bias appraisal, and an evaluation of what it changes for the people who use it. None has been done. 4 criteria below are not fully met.

    Qualifying criteria

    • MetHealth conditionchecked automatically

      A source-backed description of the health condition or event is recorded separately from the option comparison.

    • MetExplicit decisionchecked automatically

      States the decision: "Whether a person with advanced dementia who has trouble eating should have a feeding tube, or continue careful hand feeding."

    • MetTarget audiencechecked automatically

      Names who this is for: "Families and others deciding for an adult with advanced or severe dementia who now eats or swallows poorly. Most of the evidence comes from US nursing home residents."

    • MetOptionschecked automatically

      2 options, including an explicit wait-or-decline option.

    • MetBenefitschecked automatically

      5 recorded statements read as supporting an option.

    • MetHarmschecked automatically

      7 recorded statements read as burdening an option.

    • Not metValues clarificationrecorded judgment

      Not assessed. An unassessed criterion is not a met one.

    Bias reduction

    • Partly metBest available evidencechecked automatically

      4 primary sources cited, each paraphrased closely. Dated currentness checks found no final newer edition of any source. That remains a negative currentness finding rather than proof that no update is underway.

    • Not metUser involvement in developmentrecorded judgment

      No patient, caregiver, or clinician took part in developing this journey. No usability or comprehension testing has been run on it.

    • MetNeutral presentation of optionschecked automatically

      Every option carries at least one statement in each direction, so no option is recorded as all upside or all downside.

    • MetComplete citationschecked automatically

      4 sources, each with a full citation, a resolving identifier, a publication date, and the date and route it was read.

    • MetPublication or version datechecked automatically

      Last reviewed 2026-11-14.

    • Not metUpdate policyrecorded judgment

      Each journey records when it was last reviewed. A recheck cadence and triggers are proposed, and source recheck due dates are computed from them, but no owner has adopted the policy. This stays unmet until one does.

    • MetFunding disclosedrecorded judgment

      Funding and editorial independence are stated at /independence.

    Screening decisions

    • Not applicableWhat the test measureschecked automatically

      This journey is not a screening decision.

    • Not applicableConsequences of a positive resultchecked automatically

      This journey is not a screening decision.

    • Not applicableNext steps after either resultchecked automatically

      This journey is not a screening decision.

    • Not applicableOverdiagnosis and false positiveschecked automatically

      This journey is not a screening decision.

    This checks the seven IPDAS 5.0 qualifying criteria, seven bias-reduction criteria, and four screening-specific criteria against what each journey holds. IPDAS 5.0 contains more than these. A criterion absent here has not been assessed and is not passed. Standard: IPDAS 5.0, published 2026-06-12. Assessment is by Values in Care, not by IPDAS or any outside reviewer.

    Sources

    Sourced
    • Davies N, Barrado-Martín Y, Vickerstaff V, et al. Enteral tube feeding for people with severe dementia. Cochrane Database Syst Rev. 2021;8(8):CD013503.

      https://doi.org/10.1002/14651858.CD013503.pub2

      Read 2026-11-14 from the PubMed record (structured abstract of the version of record).

      Currentness check · 2026-11-14

      No updated Cochrane review, and no later systematic review of the same comparison, was identified among the results reviewed.

      Method: PubMed search via the PubMed MCP connector for systematic reviews or meta-analyses of tube feeding in dementia published from September 2021.

      Not covered: Cochrane's own list of updates in progress was not checked; only the first ten results were reviewed.

      Adjacent evidence, not incorporated into this journey: A 2026 meta-review of physical interventions for people living with advanced dementia in the community (BMJ Support Palliat Care, PMID 41698802) may touch feeding; a clinical reviewer should decide whether it bears on this journey.

    • American Geriatrics Society. American Geriatrics Society feeding tubes in advanced dementia position statement. J Am Geriatr Soc. 2014;62(8):1590-1593.

      https://doi.org/10.1111/jgs.12924

      Read 2026-11-14 from the PubMed record (structured abstract of the version of record).

      Currentness check · 2026-11-14

      The 2014 position statement is the only such record indexed; no later version was identified.

      Method: PubMed search via the PubMed MCP connector for American Geriatrics Society corporate-author records on feeding or tubes in dementia.

      Not covered: The society's own website and guideline list were not checked.

    • Teno JM, Gozalo PL, Mitchell SL, et al. Does feeding tube insertion and its timing improve survival? J Am Geriatr Soc. 2012;60(10):1918-1921.

      https://doi.org/10.1111/j.1532-5415.2012.04148.x

      Read 2026-11-14 from the PubMed record (structured abstract of the version of record).

      Currentness check · 2026-11-14

      An observational cohort is not superseded by later studies; later work adds to it rather than replaces it.

      Method: PubMed record review via the PubMed MCP connector.

      Not covered: Corrections or later reanalyses of this cohort were not searched for separately.

    • Teno JM, Gozalo P, Mitchell SL, et al. Feeding tubes and the prevention or healing of pressure ulcers. Arch Intern Med. 2012;172(9):697-701.

      https://doi.org/10.1001/archinternmed.2012.1200

      Read 2026-11-14 from the PubMed record (structured abstract of the version of record).

      Currentness check · 2026-11-14

      An observational cohort is not superseded by later studies; the 2021 Cochrane review includes evidence on the same outcome.

      Method: PubMed record review via the PubMed MCP connector.

      Not covered: Corrections or later reanalyses of this cohort were not searched for separately.

    Each clinical statement above paraphrases one of these sources closely. Filing it under a priority, or reading it against priorities you saved, is a Values in Care editorial operation rather than a source finding.

    Open Source Watch