Palliative Care

This document is a clinical practice guideline, not an experiment. It does not test an intervention. Instead, it defines:

**The intervention:** A multi-component palliative care program delivered concurrently with cancer treatment, including symptom management (pain, dyspnea, nausea, fatigue, insomnia, delirium), psychosocial and spiritual support, advance care planning, and coordination of care by an interdisciplinary team (physicians, nurses, mental health professionals, chaplains).

**The comparator:** No comparator is tested. The guidelines assume palliative care should be integrated from the initial visit for all cancer patients.

**Outcome measures:** The guidelines list desired outcomes — "satisfactory" response includes adequate pain and symptom control, reduced patient/family distress, acceptable sense of control, relief of caregiver burden, strengthened relationships, optimised quality of life, personal growth, and a "good death." But no specific measurement instruments, thresholds, or validated scales are mandated.

**No participants were studied.** This is a consensus document developed by an interdisciplinary panel of 18 experts from NCCN member institutions, including medical oncologists, neurologists, psychiatrists, anesthesiologists, internists, palliative care specialists, and geriatric medicine specialists. The guidelines are intended for all cancer patients, regardless of disease stage or treatment intent. The document does not report any sample size, demographic data, or recruitment criteria because it is not a research study.

**No measurements were taken.** The guidelines recommend screening for palliative care needs at the initial visit and at appropriate intervals, using criteria such as:

Uncontrolled symptoms

Moderate to severe distress related to cancer diagnosis or therapy

Serious comorbid physical and psychosocial conditions

Life expectancy ≤ 12 months

Poor performance status (ECOG ≥ 3 or KPS ≤ 50)

Specific clinical indicators (hypercalcemia, brain metastasis, delirium, spinal cord compression, cachexia, malignant effusions, bilirubin ≥ 2.5, creatinine ≥ 3)

However, the guidelines do not specify which validated instruments to use for screening or assessment. They reference other NCCN guidelines for specific symptoms (e.g., Distress Management, Cancer-Related Fatigue, Delirium Interventions) but do not provide the actual scales or cut-off scores.

**Study design:** This is a clinical practice guideline developed through expert consensus, not a research study. The NCCN uses a formal process: panel members review existing evidence (primarily from randomised controlled trials, systematic reviews, and meta-analyses), discuss recommendations, and vote on the strength of each recommendation. Recommendations are assigned categories:

**Category 1:** Based on high-level evidence (e.g., RCTs) with uniform consensus

**Category 2A:** Based on lower-level evidence with uniform consensus (default for most recommendations here)

**Category 2B:** Lower-level evidence, nonuniform consensus but no major disagreement

**Category 3:** Any level of evidence, major disagreement

**What this design can and cannot prove:**

**Can prove:** Nothing directly. Guidelines synthesise existing evidence but do not generate new data. They represent expert opinion about best practices.

**Cannot prove:** Causality, comparative effectiveness, effect sizes, or superiority of one approach over another. The document does not report any statistical analyses, confidence intervals, p-values, or effect sizes because it is not a study.

**Major methodological weaknesses:**

**No original data:** Zero participants, zero measurements, zero statistical tests.

**No systematic review methodology:** The document does not describe how literature was searched, selected, or synthesised. No PRISMA flow diagram, no risk of bias assessment, no meta-analysis.

**No blinding or randomisation:** Not applicable — this is not an experiment.

**Industry funding disclosures:** Panel members disclosed financial support from industry, but the document only states that disclosures were collected — it does not report what those disclosures were (individual disclosures are listed on page 473, which is not included in the provided text).

**Consensus-based, not evidence-based:** Most recommendations are Category 2A, meaning they are based on "lower-level evidence" (expert opinion, case series, non-randomised studies) with uniform consensus. This is a weak evidence base.

**No update schedule specified:** Guidelines can become outdated; the document does not state when it will be revised.

**There are no findings.** This document contains no results from any experiment, trial, or observational study. The "findings" are expert recommendations:

All cancer patients should be screened for palliative care needs at their initial visit, at appropriate intervals, and as clinically indicated.

Patients and families should be informed that palliative care is an integral part of comprehensive cancer care.

Palliative care should be delivered by an interdisciplinary team.

Skilled palliative care specialists should be readily available for consultation or direct care.

Palliative care can be delivered concurrently with life-prolonging care or as the main focus of care.

Criteria for early consultation with a palliative care specialist include: uncontrolled symptoms, moderate-to-severe distress, serious comorbid conditions, life expectancy ≤ 12 months, poor performance status (ECOG ≥ 3 or KPS ≤ 50), and specific clinical indicators (hypercalcemia, brain metastasis, delirium, spinal cord compression, cachexia, malignant effusions, elevated bilirubin or creatinine).

For imminently dying patients, interventions include: response to requests to withdraw or withhold life-sustaining treatment, response to requests for physician-assisted suicide and euthanasia, care of the imminently dying patient, and palliative sedation.

After death, interventions include: general support for family and caregivers, after-death support, immediate after-death care, and bereavement support.

**No effect sizes, confidence intervals, p-values, or numbers needed to treat are reported anywhere in the document.**

**Not applicable.** There are no effect sizes to report. The document does not quantify the benefit of palliative care compared to usual care, alternative interventions, or no intervention. It does not report:

How much pain scores decrease (e.g., "average 2-point reduction on 0–10 scale")

How much quality of life improves (e.g., "5-point increase on FACT-G")

How many patients achieve adequate symptom control (e.g., "60% vs. 40%")

Survival differences (e.g., "median survival 2.7 months longer")

Number needed to treat for any outcome

Without these numbers, a reader cannot judge whether the recommended interventions are worth the time, cost, or effort.

**What the authors acknowledge:**

The guidelines are a "statement of consensus" and "any clinician seeking to apply or consult these guidelines is expected to use independent medical judgment."

Recommendations are based on "lower-level evidence" (Category 2A) unless otherwise noted.

The guidelines are copyrighted and may not be reproduced without permission.

**What a critical reader would note:**

**No original data:** This is not a research study. It cannot be used to estimate effect sizes, compare interventions, or determine causality.

**No systematic review:** The document does not describe how evidence was gathered, evaluated, or weighted. There is no mention of databases searched, inclusion/exclusion criteria, or quality assessment of included studies.

**Consensus bias:** Expert panels can be influenced by groupthink, dominant personalities, and financial conflicts of interest. The document does not report individual disclosures (they are on page 473, not provided).

**No patient involvement:** Patients and families were not included in the guideline development process. The guidelines are written by clinicians for clinicians.

**No implementation guidance:** The guidelines do not specify how to implement screening, what instruments to use, what cut-off scores trigger referral, or how to measure outcomes.

**No cost-effectiveness analysis:** The guidelines do not address whether the recommended interventions are cost-effective or feasible in resource-limited settings.

**Population limits:** The guidelines are specific to cancer patients. They may not apply to other life-threatening or debilitating illnesses.

**Duration:** The guidelines do not specify how long palliative care should continue, how often to reassess, or when to transition to hospice.

**No comparison to usual care:** The guidelines assume palliative care is beneficial, but they do not present evidence comparing palliative care to standard oncology care alone.

**For someone running their own n=1 experiment:**

This document is not suitable as a basis for a self-experiment. It provides no specific interventions, doses, durations, or measurement tools. However, if you want to test whether adding palliative care principles to your own health management improves quality of life, here is what you would need to do — but note that the guidelines do not provide the necessary specifics:

**What to test (specific intervention and dose):**

You cannot test "palliative care" as a single intervention because it is a multi-component program. You would need to pick one component: e.g., weekly 30-minute sessions with a therapist focused on advance care planning, or daily symptom tracking with a nurse, or a structured spiritual care program.

The guidelines do not specify dose (frequency, duration, intensity) for any component.

**Minimum meaningful duration:**

The guidelines do not specify. For symptom management, a reasonable minimum might be 4–8 weeks to see changes in pain, nausea, or fatigue. For psychosocial outcomes, 8–12 weeks may be needed.

For advance care planning, a single session may be sufficient to document preferences, but the impact on quality of life may take longer to manifest.

**What to measure (specific metrics):**

The guidelines list outcomes but do not specify instruments. You would need to choose validated scales:

- Pain: 0–10 numeric rating scale (daily)

- Quality of life: FACT-G (Functional Assessment of Cancer Therapy – General) or EQ-5D

- Distress: NCCN Distress Thermometer (0–10)

- Fatigue: FACIT-Fatigue Scale

- Anxiety/Depression: HADS (Hospital Anxiety and Depression Scale)

Measure at baseline, weekly during intervention, and at follow-up.

**Key confounds to control for:**

**Concurrent treatments:** Cancer treatment (chemotherapy, radiation, surgery) changes over time and affects symptoms. Track all treatments and their timing.

**Disease progression:** Worsening cancer can increase symptoms regardless of palliative care. Monitor disease status (e.g., tumor markers, scans).

**Placebo effect:** Any new intervention can produce temporary improvement due to attention and expectation. Use a run-in period (2 weeks of baseline tracking) and consider a crossover design if ethical.

**Regression to the mean:** Symptoms fluctuate naturally. Track for at least 2 weeks before starting the intervention to establish a stable baseline.

**Co-interventions:** Pain medications, antiemetics, antidepressants, and other drugs may change during the study. Keep a medication diary.

**Social support:** Changes in family support, financial situation, or living arrangements can affect outcomes independently.

**Mood and sleep:** Depression and insomnia can amplify symptom perception. Measure these as covariates.

**What a positive result would look like:**

A ≥2-point reduction on a 0–10 pain scale (minimally clinically important difference for cancer pain)

A ≥5-point increase on the FACT-G (minimally important difference for quality of life)

A ≥2-point reduction on the Distress Thermometer

A ≥3-point reduction on the FACIT-Fatigue

Improvements should be sustained for at least 2 consecutive weeks and exceed the variability seen during the baseline period (e.g., the change should be larger than 2 standard deviations of baseline measurements)

**Bottom line:** This guideline document is useful for understanding what expert clinicians recommend for cancer patients, but it is useless for designing a self-experiment. You would need to find the original studies that the guidelines are based on — and those studies would provide the specific numbers, effect sizes, and protocols needed to run your own n=1 trial.