Abstract
Objectives:
This study was conducted before an evidence review on Traditional and Complementary Medicine (TCM) to update the clinical practice guidelines (CPGs): “Deciding palliative and end-of-life (P/EoL) care for people with diabetes.” The aim was to frame the PICO (population/problems, interventions/comparisons, and outcomes), ascertain their importance, and identify other modifying factors for grading recommendations.
Design:
A systematic scoping review mapped information about diabetes P/EoL problems and outcomes, TCM use, provision, benefits and risks, and stakeholder preferences and values. Thirteen electronic databases were searched in 2017/18 until no new information was identified. Relevant data were extracted, rated for quality, directness, and relevance, and synthesized using triangulation methods. Excluded was diabetes prevention or treatment, as this is not an important P/EoL problem.
Results:
Of the 228 included articles, except for diabetes P/EoL problems, insufficient direct evidence led to data being extrapolated from either adults with diabetes or any P/EoL diagnosis. The findings affirmed that caring for people with diabetes in need of P/EoL care is complex due to multiple fluctuating needs that are influenced by the P/EoL trajectories (stable, unstable, deteriorating, terminal, or bereaved), multimorbidity, and difficult-to-manage chronic and acute problems. The only problem specific to diabetes P/EoL care, was unstable glycemia. Over 50 TCM interventions commonly used by patients and/or provided by services were identified, of which, many might simultaneously address multiple problems and 18 had been appraised in systematic reviews. Physical and psychologic symptom reliefs were most often evaluated; however, these were only one aspect of a “good death.” Other important outcomes were the quality and location of care, personal agency, relationships, preparations for the dying process, spirituality, and affirmation of the whole person. Other important modifying factors included opportunity costs, affordability, availability, preferences, cultural appropriateness, and alignment with beliefs about the meaning of illness and death.
Conclusions:
There is a role for TCM in the multidisciplinary holistic P/EoL care of people with diabetes. Due to the paucity of evidence specific to this population, the generalizability of some of these results is broader and the updated CPG will also need to consider indirect evidence from other patient groups. Along with recommendations about indications for TCM use, the CGP should provide guidance on ceasing unnecessary interventions, reducing polypharmacy and managing unstable glycemia is required. Before ceasing a TCM, a broader risk-benefit analysis is recommended, as unlike many conventional therapies, there may be multiple benefits warranting its continuation.
Introduction
Diabetes mellitus is a common palliative and end-of-life condition due to its increasing prevalence and associated complications. In 2016, diabetes was the seventh most common indirect cause of death globally, 1,2 and associated with direct causes of death, such as cancer, coronary heart disease and stroke. 3 Chronic diabetes complications that continue until the end of life, such as kidney disease, neuropathy, foot ulcers, lower limb amputation, and vision impairment, account for over 50% of the burden of disease for people with diabetes in Australia, aged 75 years or more. 4
Palliative care includes but is not limited to end-of-life care. Although death can be sudden, the trajectory for many people with diabetes is often long with numerous periods of instability, complicated by multimorbidity. 5 For others, diabetes may only develop once they are palliative, due to the introduction of medications such as corticosteroids for symptom management of another condition, such as cancer. 6 Palliative/end-of-life care aims to be holistic and person-centered, emphasizing quality-of-life and the relief of suffering over curative treatments and invasive management. 5 Diabetes management decisions are often complex, 6 –9 highlighting the importance of tools such as clinical practice guidelines (CPGs) that provide relevant up-to-date evidence-based guidance for patients, caregivers, health providers, and policy makers.
To this end, the 2010 Australian Guidelines for managing people with diabetes at the end of life was due to be updated. 6 The review was expanded from end-of-life care to include all aspects of palliative care. Thus, in alignment with the CPG, the term palliative or end-of-life (P/EoL) care is used throughout the article.
The substantial use of Traditional and Complementary Medicine (TCM) by this patient group, 10,11 emerging evidence of TCM benefits, 12,13 potential for interactions with conventional medical management 14 and its increasing provision by palliative services 15 prompted the guideline panel to systematically review the evidence and integrate TCM recommendations into the updated guidelines. Furthermore, it was anticipated that the inclusion of TCM could potentially enrich evidence-based person-centered information presented in the guideline due to alignment of TCM and palliative care philosophies and values of patients. 16 –20
CPGs make evidence-informed recommendations based on the certainty of evidence for benefits and harms, 21 which are then qualified by modifying factors such as the burden of disease, magnitude of effect, current use, patient values and preferences, and equity. 22 However, it is not uncommon for CPGs to inadequately frame the PICO research questions (i.e., population, interventions, comparisons and outcomes) for TCM interventions and instead, focus solely on evidence statements, ignoring other modifying factors relevant to TCM that may warrant upgrading or downgrading the strength of recommendations. 23
Furthermore, similar to many CPGs, 24,25 the consideration of TCM both in the 2010 Australian version 6 and comparable diabetes P/EoL guidelines from the United States, 26 Canada, 27 and United Kingdom 28,29 appears to be ad hoc and often provides no recommendations for or against TCM use.
In response, the guideline panel for the updated Australian “Guidelines for deciding palliative and end-of-life care for people with diabetes” 30 aimed to conduct a transparent and systematic approach to the inclusion of TCM recommendations in the CPG through a two-stage approach:
Stage 1: A scoping review to inform the framing of the PICO research questions for the evidence review in Step 2 and identify modifying factors for qualifying recommendations when updating the CPG (Table 1).
Mapping Information About Traditional and Complementary Medicine Use, Provision, Benefits, and Risks
TCM, Traditional and Complementary Medicine.
Stage 2: A rapid evidence review evaluating the benefits and risks of up to 20 TCM interventions for the updated CPG.
This article presents the results of Stage 1: A scoping review.
Materials and Methods
The systematic scoping review was conducted using the methodology outlined by Levac et al. 31 and reported according to the PRISMA-ScR extension. 32
Protocol
A protocol for the review was published in 2017. 33
Eligibility
Included were publications reporting information about the problems experienced by adults with diabetes mellitus in need of P/EoL care and/or their caregivers. Also included were publications reporting information about TCM use, service provision, benefits or risks, important P/EoL outcomes, and other modifying factors relevant to adults with diabetes or adults with any diagnosis in need of P/EoL care (Table 1). Peer review articles and gray literature published in English were included. There were no limits on publication date.
Excluded were preclinical, laboratory, and nonhuman studies and studies measuring nonclinical or nonpatient-reported outcomes. Studies evaluating TCM for diabetes prevention, risk reduction, or treatment were also excluded as these are not important P/EoL problems. The exception was studies reporting information directly relevant to managing unstable glycemia, as this is a common P/EoL diabetes problem. 34 Publications rated 1 (see Rating) were also excluded.
Search
Publications were sought from 13 electronic databases, bibliographies, and cluster searching of key authors using an iterative process until data saturation was achieved. Three searches were conducted on December 22, 2017, January 4, 2018, and September 10, 2018. The first two searched for any publication type with a focus on identifying TCM use, patterns of use, and risks. The third searched for systematic reviews of effectiveness of TCM interventions and for additional information about common problems or important outcomes. Appendix 1 provides further details of the database searches and keywords.
Publication selection
Two authors (S.A. and J.H.) independently screened the title/abstract. One author (S.A. or J.H.) screened the full texts, referring to the other authors for further discussion and consensus. Ten percent of full texts were randomly selected and independently screened by another author. Disagreements were resolved through consensus.
Data extraction
Electronic outputs were downloaded, merged, and stored in the electronic citation manager Endnote, 35 from which data were extracted into electronic spreadsheets. Data extraction included information about the study design, participants, P/EoL trajectories (stable, unstable, deteriorating, terminal, or bereaved), 36 setting, problems/conditions, TCM utilization or provision and the reasons, outcome measures, other modifiers (e.g., values/preferences, cost, quality, accessibility, feasibility, and considerations for specific subgroups), and potentially effective interventions and risks (e.g., safety, interactions, pharmacokinetics, hypo/hyperglycemia, and contraindications).
Rating
Data were scrutinized at all levels of the extraction process and analysis. For each category, publications were rated on a scale of 1–5, with 5 reflecting higher certainty/quality and directness of the evidence. CASP 37 and AGREE II 38 checklists were used when rating quality. At least two authors independently rated each publication.
Publications evaluating effectiveness or risks were first rated according to the quality of the evidence. 37,39 A rating of 5 was limited to Level 1 studies as defined by the Oxford Centre for Evidence-Based Medicine. 40 A maximum rating of 4 was given to Level 2 studies, and so on for the 5 Levels. From there, publications were rated up or down according to quality, directness of the evidence to people with diabetes at the end of life, their caregivers and P/EoL health care settings in Australia.
The opposite approach was used when rating all other categories, that is, the publication was first rated according to the directness and generalizability of the evidence, followed by quality.
For instance, publications were given a provisional rating of 5 if participants or the health setting was Australian, rated 4 when from New Zealand, United Kingdom, United States, and Canada, and rated 3 for the most common culturally and linguistically diverse (CALD) groups in Australia (i.e., India, China, Hong Kong, Vietnam, Malaysia, Philippines, Netherlands, Germany, Italy, Greece, and Lebanon) or conducted in a European health setting. Publications in the outcomes category were also rated higher if the study directly reported stakeholders' views about the importance of an outcome or the psychometric properties of outcome measurement tools for this population, and lower for indirect evidence that most commonly extracted from effectiveness studies. 41 From there, ratings were further adjusted according to the certainty of the evidence.
Synthesis of results
The principle measures were descriptive. Synthesis drew on triangulation methods. 42 The tabulated results and publication ratings were mapped against the PICO-MER (Table 1) and interrogated for convergent, additional, or contradictory information. Publications with lower ratings were only included in the final narration if data saturation was not achieved.
Results
The search identified 888 citations, from which 362 were shortlisted for full article review and 228 were included in the scoping review synthesis (Fig. 1). 5,9 –11,18,20,27,29,34,36,43 –260 Most studies reporting outcomes not relevant to people with diabetes in need of P/EoL care were excluded during the title/abstract screen. Nevertheless, this was still the most common reason for full article exclusion (n = 55), followed by populations without a diagnosis of diabetes that were not P/EoL (n = 43).

Flowchart of study retrieval and selection.
Characteristics of publications
A broad range of publications and study designs were included (Table 2) of which 13 individual studies were conducted in Australia. 10,18,78,80,81,97,98,108,111,115,141,142,176,183,185,194,238,239,249,252 A large amount of quantitative and qualitative data were available regarding prevalence of TCM use, reasons for use, and patient preferences (Table 3). Information gaps were research investigating adults with diabetes at the P/EoL stage and CALD subgroups, including no information about Australian Aboriginal and Torres Strait Islanders, TCM provided by Australian practitioners and P/EoL services, and the specifics of TCM products most often used. Additional indirect information was, therefore, extrapolated from research of adults with diabetes, with any diagnosis in the P/EoL stage or living in countries representing most common CALD groups in Australia, P/EoL services in the United States, United Kingdom, and Canada, and interventions appraised in systematic reviews or discussed in nonsystematic review articles. Most of the data used in the synthesis came from publications rated 5 or 4 (Table 3). Eighty-two publications informed two or more research questions.
Characteristics of Publications
Bold indicates significance of studies including Australian participants.
One mixed-method study included adults with diabetes and diabetes health services. 108
P/EoL, palliative and/or end of life.
Ratings of Included Publications
Multiple categories and ratings may apply to a single publication.
Relevance to diabetes, P/EoL, Australian health care settings, and directness of the evidence were most important for rating.
Certainty of evidence and level of evidence were most important for rating.
Rated 5: Highly relevant, direct evidence, and/or high-quality evidence.
Rated 4: Relevant, some direct evidence, and/or moderate-quality evidence.
Rated 3: Moderately relevant, mostly indirect evidence, and/or low-quality evidence.
Rated 2: Slightly relevant, indirect evidence, and/or low-quality evidence.
Population
The care needs of people with diabetes in need of P/EoL care are often multiple, complex, and dynamic and cross-clinical, functional, psychologic, social, and spiritual domains. 9,10,29,36,113,123,150,230 Priorities of care may change depending on the palliative trajectory (stable, unstable, deteriorating, terminal, or bereaved 36 ) and location of health care service (home, hospice, or hospital). 52 This was reflected in predictors of TCM use that included a longer duration or severity of diabetes, number of comorbidities, and declining quality of life (QoL). 9,10,29,36,52,113,123,150,183,185,199,230,239 Another important, yet perhaps less obvious finding was P/EoL care populations also includes caregivers, family members, and the bereaved. 9,36
The only problem specific to diabetes P/EoL care was unstable glycemia. Other common diabetes P/EoL problems include diabetes complications from micro- and macrovascular disease, along with less specific problems such as neuropathies, pain, nausea, anorexia, pruritis, constipation, infections, mental health, frailty, and cachexia. 9,11,29 P/EoL management challenges include breathlessness, fatigue, delirium, nonpharmacologic analgesia, complex care needs, frailty, and dementia. 150 Problems are often interrelated and exacerbated by comorbidities. For instance, anorexia, nausea and vomiting may cause hypoglycemia and exacerbate cachexia, weakness and frailty. 11 Common difficult-to-treat symptoms and problems may have multiple etiologies, such as diabetic foot disease being caused by vascular disease, peripheral neuropathy, and infection, neuropathic pain caused by diabetes and nutritional deficiencies, constipation caused by diabetic autonomic neuropathy and opiate side effects, and fatigue caused by hyper- or hypoglycemia, nutritional deficiencies, malignancy, depression, and/or dementia. 29
Polypharmacy, which includes the concurrent use of TCM products, was identified as another common problem, 10,48,86,238 as was suboptimal nutritional status such as low Vitamin B12 that is associated with metformin use. 88,178 Practitioners requested more guidance on deciding when to withdraw life-sustaining treatment and supporting patients and caregivers through the process. 9
The chronicity of diabetes and its serious complications often leave patients feeling powerless and disheartened. 252 This is further exacerbated in the P/EoL stage due to declining functioning, increasing intensity of pain, reduced capacity for self-care, and increasing psycho-social needs that in turn negatively impact physical health, QoL, and caregiver burden. 36,150,261 Spiritual concerns were also common and often insufficiently addressed by P/EoL services. 78,150 Unmet spiritual needs can manifest as physical and emotional symptoms and are associated with a higher use of health care and social services. 113,119
Interventions
Rates of TCM use worldwide by adults with diabetes ranged between 17% and 73%, 86 and were comparable with use in Australia where 48%–61% of adults with diabetes used TCM at some stage in their life and 38% used TCM in the previous 12 months. 10,238 Similar utilization rates were found for Australians at the P/EoL stage. Two small studies of adults with advanced cancer found 32%–64% were using TCM. 97,115 In the United States, 33% of adults with moderate–severe pain in hospice settings 96 and 54% in the community 244 reported using TCM.
TCM products such as vitamins, minerals, and herbal medicines were most often used both in Australia and comparable countries with rates ranging between 16% and 45%. 10,44,86,97,98,109,115,125,183,185,239 Low consultations rates with practitioners trained to prescribe TCM products (∼8%–10% of TCM users) 97,115,183 suggested substantial self-prescribing. Detailed information about the different types of TCM products used was limited and inconsistently reported. 97,98,115 The most specific information came from studies reporting ingredients commonly recommended by naturopaths for patients with diabetes (see footnote Table 5). 64,71,86,93,185
Shortlist of Interventions for an Evidence Review
Based on reasons for TCM use and outcomes domains assessed in evaluation studies.
Number of studies is an estimate as a search for RCTs and NRSIs was not conducted; evidence of effectiveness, safety, or costs was not appraised.
Limited information identified about individual or combination ingredients.
e.g. Reiki, therapeutic touch, Pranic healing, and kinesiology; prayer and religious interventions were excluded as these are not classified as TCM in Australia.
Individual or combination Western, Chinese, Ayurvedic, or other traditional herbs and supplements commonly used by lay persons, prescribed by practitioners or recommended in nonsystematic review articles: Allium sativum, Astragalus membranaceus, Camellia sinensis, Cinnamomum sp., Curcuma longa, Eleutherococcus senticosus, Ginkgo biloba, Gymnema sylvestre, Nigella sativa, Olea Europa, Panax ginseng, Panax quinquefolius, Plantago psyllium, Rehmannia glutinosa, Trigonella foenum-graecum and Zingiber officinalis. Alpha-Lipoic acid, antioxidants, CoEnzyme Q10, carnatine, chromium, N-acetyl-cysteine, magnesium, omega3, vitamin B's, C, D3 and E, and zinc.
NRSIs, nonrandomized studies of interventions; P/EoL, palliative and end of life; RCTs, randomized controlled trials; SRs, systematic reviews; TCM, Traditional and Complementary Medicine; TENS, transcutaneous electric nerve stimulation.
In Australia the most common consultations were with energetic healers (e.g., Reiki, therapeutic touch, and kinesiology), massage therapists, naturopaths, acupuncturists, integrative medical practitioners, and music or art therapists. 10,97,98,115,183 Other commonly used TCM included movement therapy (e.g., yoga, qigong, and t'ai chi) and meditation or relaxation. Combinations were often used such as massage, aromatherapy, and music 97 or multiple vitamin mineral and herbal products. 10,97
There was a paucity of Australian health services research. Excluding religious support and pastoral care, studies from the United States, Canada, and United Kingdom suggest that noningestible TCM such as massage/therapeutic touch, relaxation/meditation, music/pet/art therapy, and acupuncture/acupressure are most commonly provided by P/EoL services. 60,102,170,208,209,250
Thirty-one systematic reviews evaluated the effectiveness of TCM interventions for 18 interventions (Table 4). 43,47,66,72,82,89,92,116,117,119,121,126,134,144,145,159,162,165,166,174,190,196,212,218,222,226,231,243,245,256 Two were empty reviews of TCM evidence 196,222 and as such, are not presented in Table 4. Other high-quality systematic reviews for P/EoL care included one realist review of music therapy, 189 one meta-synthesis of spiritual interventions, 113 one scoping review of nonpharmacologic interventions, 95 and one narrative review of TCM interventions. 254
Systematic Reviews of Effectiveness
The outcomes listed do not confirm effectiveness, as risk of bias and certainty of the evidence were not appraised.
P/EoL, palliative and/or end of life; QoL, quality of life; TCM, Traditional and Complementary Medicine; TENS, transcutaneous electric nerve stimulation.
Many of these interventions, along with numerous others that were not reviewed systematically, have the potential to simultaneously address multiple problems or symptom clusters. 11,67,83,154,171 For instance, acupuncture, acupressure, and/or Chinese herbal medicine may simultaneously assist neuropathic pain, fatigue, dyspnea, insomnia, nausea cognitive function, and QoL. 67,134,166,173,174,212,234,256,259 Massage may simultaneously reduce pain, anxiety, depression, and nausea, provide comfort and improve coping, 95,106,107,121,160,212,219 and combined with aromatherapy, may additionally reduce agitation and improve well-being. 74,95 Music therapy may alleviate a range of physical and psychologic symptoms, provide comfort and relaxation, improve coping, support socialization, and uplift the spirit. 94,95,113,132,137 –141,189,220,231 Art therapy may relieve pain, anxiety, depression, and improve fatigue and well-being. 95 Biofield therapies, such as Reiki or therapeutic touch, may relieve pain, reduce anxiety, and improve QoL. 76,77,137 Spiritual/religious interventions may alleviate spiritual distress, anxiety, depression, and chronic pain. 82,119
Naturopathy may alleviate physical and psychologic symptoms and support self-efficacy and motivation. 68,70,71,252 Examples of simultaneous effects of herbs or nutritional supplements include Curcuma longa (turmeric) and Zingiber officinalis (ginger) for musculoskeletal pain relief, dyspepsia, and nausea 154,171,184 and zinc for mood disorders, periodontal disease, infections, neuropathy, and retinopathy. 145,222
Numerous reviews discussed the potential for many commonly used TCM product ingredients to interact with pharmaceuticals or affect glycaemic control, often information about potential benefits was also presented. 64,131,167,175,187,234,237,247,260 The risks of polypharmacy and concurrent use of TCM products were also discussed, mostly in the context of diabetes 48,80,86,167,238 or as elderly patients receiving palliative care. 171 Other publications reviewed the potential risks and benefits associated with TCM product use, 168,171,184,257 acupuncture, 122,240 moxibustion, 213 or massage. 217
Table 5 summarized the availability of different types of evidence proposed in Table 1 for shortlisting up to 20 TCM interventions for the CPG evidence review. The findings complement and build on the scoping review of nonpharmacologic interventions for P/EoL comfort that recommended conducting systematic reviews of effectiveness for massage therapy and music therapy. 95
Outcomes
A broad range of holistic health outcomes were identified as important to adults with diabetes. Bar one study, 185 rather than treating diabetes, the primary reasons for TCM use by adults with diabetes in Australia and the United States, including older adults in the United States, was to manage acute or chronic symptoms, prevent illness, and improve general health, well-being, and QoL. 18,48,57,73,81,108,176,183,199,252 Similarly, common reasons for TCM use by adults during the P/EoL stage were symptom relief, immune system support, well-being, and QoL. 98,115,244 Seeking empowerment, autonomy, agency, and self-efficacy were important nonclinical reasons for TCM use. 11,87,98,201,252
Comfort and peacefulness were repeated in the literature as being valued by adults in the P/EoL stage and their caregivers 56,78,94,95,98,106,127,135,201,215,217,242,244 and relief of physical and psychologic suffering, although important, was only one aspect of a “good death.” 106,242 Bar one study where medical doctors generally thought physical symptom relief was most important, 242 patients, caregivers, and practitioners considered other process-oriented and experiential outcomes to be equally as important. 56,78,98,115,242 These outcomes were concerned with quality of care, patient-reported experiences, their satisfaction with care, self-determination and control, preparations for the dying process and death, place of death, relationships, ease of loneliness, spirituality, and affirmation of the whole person. 53,56,61,62,81,95,108,113,115,119,141,220,242,252 As people moved through the P/EoL stages, process-oriented and experiential outcomes became increasingly important 61,78,107,115,201,242 as did the outcomes and experiences of caregivers, family, and the bereaved. 53,103 In comparable countries such as Canada, United States, and United Kingdom, TCM services aimed to improve patient, family, and caregiver experiences with P/EoL care 62,99,103,195,209,216,250 and in some instances improve economic viability. 99
Despite the importance of experiential outcomes, the primary outcomes in most systematic reviews of TCM for P/EoL care were physical or psychologic symptom relief, followed by improving QoL and functioning. 82,92,121,162,166,190,196,212,223,226 It is, therefore, possible that outcomes most important to patients were not evaluated, as was demonstrated by a nonsolicited comment from a study participant in a randomized controlled trial that demonstrated no significant improvement in QoL nor pain from massage or meditation for P/EoL care. 107
You never ask how important it is to me to receive this service. .. I so much look forward to it. .. I'm amazed you don't ask me this question. It should be the featured question. 107
Similarly, although spiritual outcomes were important to patients, few systematic reviews listed it as a primary outcome. 72 This included a Cochrane review of interventions with a spiritual or religious component for adults at the end of life. 82 Instead, spiritual well-being, along with physical symptom relief, health service use, place of death, and caregiver outcomes were secondary outcomes. Another Cochrane review of music therapy initially included spiritual outcomes, 72 only to remove it from the updated review. 190 Examples of nonreligious spiritual outcomes in the P/EoL setting are relationships with self and others, hope, meaning and purpose in life, being present, “in tune with life” and at peace. 113,119,201,217 Indeed, many TCMs have the potential to positively impact these outcomes 56,62,76,94,106,141,151,162,201,217 and some aspects can be measured using validated outcome tools. 72,113,119 Notwithstanding, there are many challenges with assessing spiritual outcomes and one systematic narrative review noted ongoing debate about whether evidence-based medicine can even be applied to this domain. 113
In contrast to studies evaluating TCM effectiveness, studies evaluating P/EoL services with or without TCM, generally used a broader range of outcomes and often considered caregivers' experiences 36,52,61,103,172,195 For instance, a systematic review of studies evaluating in-home P/EoL programs found the outcomes most often assessed were patient functioning and symptom relief, patient and/or caregiver QoL or satisfaction, place of death, and health service use/costs. 52
Regarding outcome measurement in clinical trials, mostly studies used validated instruments, such as clinical measurements, laboratory investigations, and patient-reported outcome measures (PROMs). Concerns were raised, however, that not all PROMs have been validated for use or are sensitive to change in the P/EoL setting. 94,107
Although PROMs and patient experience outcome measures (PREMs) are generally preferred, it is not always possible due to deteriorating health and responder burden. As such, some evaluation studies used a combination of patient and caregiver/practitioner-reported measures, 47,157,188 which have been implemented by the Australian Palliative Care Outcomes Collaboration (PCOC). 36
Modifying factors
Demand and uptake of available TCM varied widely, which in part was due to different study samples, questionnaires, and definitions of TCM. TCM was mostly used as an adjuvant rather than an alternative to conventional medicine, 18,86,97,98,115,176 suggesting medical opportunity costs were low. Demand was predicted by chronicity and severity of disease, a higher number of comorbidities and declining health status. 114,199,238
The P/EoL stage and location (e.g., home vs. institutional care) were also found to influence uptake and referrals, including self-referrals. 141,180
Decisions to use TCM reflected a preference for “natural things,” safer therapies with fewer side effects, self-determination, and self-care and/or seeking a better health care experience, particularly person-centered care. 18,80,97,98,115,176 Although clinical effectiveness was important, patients and caregivers also highly valued holistic approaches that supported a mind–body connection and well-being. 78,81,97,98,115,203,252 P/EoL patients and caregivers valued the bidirectional benefits they experienced from TCM. 77,103,201,244 Patients often wanted professional advice about TCM use and stated that a doctors' recommendations influenced their decisions. Often, however, TCM was self-prescribed or recommended by family or friends, with few receiving advice from appropriately qualified TCM or conventional health care practitioners. 10,108,109,176,183 Opportunity costs of not accessing potentially effective TCM included increasing pain, anxiety, and lower satisfaction with health care services. 97,103
Results were mixed about whether ethnicity, country of birth, and language spoken at home are predictors of TCM use in Australia or the United States. 48,57,185,239,244 Notwithstanding, for some CALD subgroups, TCM was used because it aligned with beliefs about the meaning of illness and death, culturally appropriate approaches for health care, or the importance of individual and/or community responsibility for health and dying. 59,85,104,119,130,182,185,215,236 Notably, no information was identified pertaining to TCM use, preferences, and values of Australian Aboriginal and Torres Strait Islanders.
Affordability further influenced use, raising concerns about equity. 97,176,183,209 Higher socioeconomic status was a consistent predictor of use in Australia 10,97,183,185 and comparable countries. 57,244 Out-of-pocket costs was a substantial barrier for adults with diabetes 10 and at the end of life 115 accessing TCM in Australia. In 2010, median TCM expenditure by adults with diabetes in Australia was estimated at AUD$240/annum for practitioner services, and AUD$360/annum for products. 10,238 Financial barriers were also identified as the most common and largest barrier to TCM service provision, with many hospices in the United States and Canada relying partially or solely on volunteer TCM practitioners 60,99,102,156,172,207,209,250 or research grants 172 to implement services. 172
Although only four studies were identified that reported TCM service provision in Australian P/EoL. 78,141,194,249 Studies from comparable countries evaluating TCM implemented in home care, hospice, and palliative care settings, 95 suggest it is feasible. Rates ranged from 42% to 90% of outpatient, hospice, and hospital services in the United States, 60,99,102,208,250 11% of hospices in Canada (with a further 45% allowing visiting TCM practitioners) 209 and 54% of UK hospice, in- and outpatient and home services. 170 Often, however, the uptake of services was low. 60,96,103,207 Finding qualified practitioners, a lack of service know-how and the short duration of many episodes of care were other barriers to TCM provision. 60,99,102,104,108,113,156,172,207,209,210,216,250 A devaluing attitude toward TCM was also reported 11,115,133,136,147,240 that in turn negatively impacted TCM referrals or provision despite evidence of effectiveness, improved patient and caregiver satisfaction, and/or improved job satisfaction. 77,103,139,189,194,201,249
Discussion
This review mapped and synthesized a wide range of research to inform the updating of the Australian CPGs for people with diabetes in need of P/EoL care. Except for information about common and important problems, the paucity of evidence specific to the population led to the incorporation of a large amount of less direct evidence from other countries about adults with diabetes and adults with any diagnosis in need of P/EoL care. Notwithstanding, the broader inclusion criteria was necessary before conducting an evidence review and updating the CPGs. Guideline developers are encouraged to make and qualify recommendations based on the best available evidence. No recommendations, nonspecific recommendations, and recommending more research are of little help to clinicians and patients who must make decisions despite uncertainty. 262
Transparency and rigor are, therefore, important when framing PICO questions for CPG evidence reviews and considering modifying factors when qualifying recommendations. Solely relying upon the expertise of the guideline panel can unintentionally bias this process, thereby limiting the quality and usefulness of recommendations. 23,25,41 To this end, along with TCM interventions that are provided by P/EoL services, or where there is evidence about benefits and risks, CPGs should also consider interventions that are commonly used by patients or at least state why they were out of scope. 23
Over 50 individual TCM interventions were identified that have a potential role in the care of people with diabetes in need of P/EoL care. However, pragmatic constraints necessitate a maximum of 20 for systematic review to update the guidelines. 33 Selection strategies for the guideline panel might include shortlisting interventions that have the potential to simultaneously address multiple P/EoL problems or difficult-to-treat problems, which are most likely to have a large effect on outcomes critically important to patients, or are associated with higher risks or costs. 33 Although most P/EoL services were unlikely to be prescribing TCM products, given these are most commonly used by patients, more specific guidance about their use is imperative.
The need for P/EoL guidance on ceasing unnecessary interventions, reducing polypharmacy and managing unstable glycemia was identified. 34 TCM interventions that may warrant ceasing those used by patients for disease prevention, are likely to interact with antidiabetic medication and/or have potential antidiabetic properties. However, unlike pharmaceuticals that mostly have a single action, the possibility that a TCM might also reduce other diabetes P/EoL problems should be considered before cessation. CPGs would, therefore, benefit from including a broad risk-benefit analysis of selected TCM to assist decision-making. Ideally this would include information about time taken for effects to commence and/or wear off, monetary and opportunity costs, and discussion about patient and caregiver values and preferences.
Initiatives such as the International Consortium for Health Outcomes Measurement (ICHOM 263 ) and Core Outcome Measures in Effectiveness Trials (COMET 264 ) affirm the importance of evaluating outcomes important to stakeholders. A core outcome set for P/EoL is yet to be published. With the exception of spirituality and disease prevention, the standard sets proposed by ICHOM for older people and for people with diabetes were consistent with the findings from this review. 265 Interestingly, although participation in decision-making was listed as an important outcome for older people, this was the only domain with no recommended measurement tool. Measuring outcomes such self-efficacy, enablement, and spirituality are challenging; however, validated PROMs and PREMs are available, 266 including individualized instruments such as Measure Yourself, Concerns, and Wellbeing (MYCAW) that are designed to identify and quantify changes in the outcomes most important to each person. 267
Modifying factors are mostly used to qualify recommendations. 22 This information is essential when there is limited evidence for effectiveness, as is often the case for P/EoL care. 111,268,269 In such instances, rather than no recommendation, the U.S. Preventative Services Task Force suggests structuring recommendations around the burden of suffering, potential harms, including doing nothing, monetary and opportunity costs, and feasibility of introducing or ceasing an intervention. 270 Scoping reviews such as this, along with other methodologies such as network meta-analyses and realist synthesis, may, therefore, ensure the recommendations in the CPGs extend beyond nonspecific advice to discuss TCM use with patients or conduct more research. 23 –25,271
Limitations of this review include reduced generalizability due to the Australian focus. Nevertheless, research from other countries was presented and many of the Australian results were consistent with those from other countries and other international scoping reviews. 95,188 The inclusion of articles with less direct evidence was another limitation, many of which had to be considered from different perspectives and lens to answer the research questions. Although advanced cancer was the most common diagnosis in the P/EoL studies, many of the findings were generalizable as cancer is a leading cause of death for people with diabetes. Palliative medication can cause diabetes and many of the problems experienced by the populations are similar. 3 Similarly, due to the broad inclusion criteria in this scoping review, many of the findings have broader relevance to other P/EoL researchers and guideline panels.
Another limitation was that the search terms did not include frail or elderly populations. Therefore, some systematic reviews (e.g., t'ai chi for prevention of falls) and modifying factors may have been missed. This should be borne in mind when the guideline panel selects the TCM interventions for the evidence review (Stage 2).
Conclusions
The findings from this scoping review confirm a role for TCM in the multidisciplinary holistic P/EoL care of people with diabetes in Australia. Identified evidence gaps confirm the need for P/EoL research that focuses on people with diagnoses other than cancer, CALD and older adult subgroups, and home and residential care settings. In the interim, similar to this scoping review, the guideline panel updating the CPG “Deciding palliative and end-of-life care for people with diabetes” will need to consider indirect evidence from related populations and subgroups when framing the PICO research questions, conducting a TCM evidence review and qualifying recommendations. The CPG evidence reviewers should prioritize appraising TCM interventions that can simultaneously address multiple problems, information relevant to ceasing interventions, and studies that evaluate psycho-social-spiritual outcomes and health care experiences important to patients and caregivers. The modifying factors identified should help ensure the recommendations are more than just evidence statements, thereby improving their usefulness, particularly when there is uncertainty about benefits or risks.
Footnotes
Acknowledgments
The authors thank Professor Trisha Dunning AM, Deakin University and their colleagues in the Writing Group, and the Health Professional Expert Advisory Group to the Revision of the Guidelines for Managing Diabetes at the End of Life for their input during the protocol development.
Author Disclosure Statement
All three authors are members of the Writing Group to the Revision of the Guidelines for Managing Diabetes at the End of Life, Australia. S.A. is a practicing naturopath. J.H. is a general practitioner with an interest in integrative medicine. G.D. is a general practitioner and senior research fellow at the School of Public Health and Preventative Medicine, Monash University. S.A. and J.H. are affiliated with the NICM Health Research Institute, Western Sydney University.
Funding Information
As a medical research institute, NICM Health Research Institute receives research grants and donations from foundations, universities, government agencies, individuals and industry. Sponsors and donors also provide untied funding for work to advance the vision and mission of the Institute. The project that is the subject of this article was not undertaken as part of a contractual relationship with any donor or sponsor. The authors declare no competing financial interests.
