The impact of health and wellness coaching on patient-important outcomes in chronic illness care: A systematic review and meta-analysis

Abstract

Background:

Health and Wellness Coaching (HWC) may be beneficial in chronic condition care. We sought to appraise its effectiveness on quality of life (QoL), self-efficacy (SE), depression, and anxiety.

Methods:

We searched MEDLINE, EMBASE, CINAHL, PsycINFO, and Cochrane CENTRAL for randomized trials published January 2005 - March 2023 that compared HWC to standard clinical care or another intervention without coaching. We examined QoL, SE, depression, or anxiety outcomes. Meta-analysis utilizing the random-effects model was used to estimate the pooled standardized mean difference (SMD).

Results:

Thirty included studies demonstrated that HWC improved QoL within 3 months (SMD 0.62 95 % CI 0.22–1.02, p = 0.002), SE within 1.5 months (SMD 0.38, 95 % CI 0.03–0.73, p = 0.03), and depression at 3, 6, and 12 months (SMD 0.67, 95 % CI 0.13–1.20, p = 0.01), (SMD 0.72, 95 % CI 0.19–1.24, p = 0.006), and (SMD 0.41, 95 % CI 0.09–0.73, p = 0.01) Certainty in the evidence for most outcomes was either very low or low primarily due to the high risk of bias, heterogeneity, and imprecision.

Conclusion:

HWC improves QoL, SE, and depression across chronic illness populations. Future research needs to standardize intervention reporting and outcome collection.

Practice Implications:

Future HWC studies should standardize intervention components, reporting, and outcome measures, apply relevant chronic illness theories, and aim to follow participants for greater than one year.

1. Introduction

The growing prevalence of chronic conditions represents a significant challenge to the healthcare system. Currently, one in four people lives with one or more chronic conditions, rising to three in four for the population over the age of 65 [1,2]. Furthermore, those living with multiple chronic conditions (MCC) have unique challenges, such as the cumulative self-management tasks required to manage multiple diseases; [3] for this population, there is a growing consensus that additional research is needed and new clinical strategies are required [4,5].

Health and Wellness Coaching (HWC) is an intervention that has received attention as an intervention that may meet the needs of this population [6,7]. In 2013, Wolever et al. conducted an in-depth systematic review of the healthcare coaching literature to arrive at a definition of HWC: “a patient-centered approach wherein patients at least partially determine their goals, use self-discovery or active learning processes together with content education to work toward their goals, and self-monitor behaviors to increase accountability, all within the context of an interpersonal relationship with a coach.” [8]. Additionally, coaching has the fundamental underlying assumption that people have an innate capacity to grow and focuses on constructing solutions and goal attainment [9].

While this definition may seem similar to other person-centered interventions such as various forms of psychotherapy or mindfulness exercises grounded in positive psychology, there are a few important differences. First, HWC is not designed to be a “treatment” for diagnosed mental health conditions [10,11]. Coaches may certainly work with patients that have mental health conditions and patients with these diagnoses may benefit from coaching. However, the HWC scope of practice would dictate referral for patients who appear to be unstable or struggling with these conditions to another practitioner that is trained to specifically assess mental health conditions, deliver psychotherapy, and/or can prescribe medications for treatment [10,11]. Positive psychology interventions may be adjunct to traditional psychotherapy or coaching but are delivered in a variety of ways; [12] there is no requirement of these to involve consistent interaction with an individual to monitor progress and increase accountability that is required within HWC [6].

For patients with MCC, HWC may act upon the improvement of patient capacity for self-management, which has implications for health outcomes and experience of care. When individuals have limited capacity to manage emotional problems with family and friends, role and activity limitations, financial challenges, and healthcare delivery inefficiencies, they experience greater treatment burden [13]. Furthermore, in seeking to overcome treatment burden challenges, patients draw on capacity to make adaptations over time [14]. The Theory of Patient Capacity (TPC) illustrates the components of patient capacity to consider including: successful reframing of patients’ Biography, the ability to mobilize new or existing Resources, the support of patients’ Environment, experiential accomplishment of patient and life Work, and Social functioning (BREWS) [15].

There has been considerable debate regarding the appropriate term for coaching in healthcare. Terms such as “health coaching,” “life coaching,” “wellness coaching,” have all been used to describe coaching in healthcare [7,9]. Recently, the field has sought to bring clarity by standardizing the naming as “health and wellness coaching,” or HWC, and creating certifications for HWC [6]. Despite recent standardization efforts, more variation in the delivery of coaching occurs in practice. This variation includes the educational backgrounds and certifications of those delivering coaching, modalities by which coaching is delivered, frequency of coaching sessions, and extent to which coaching is patient-centered [8,16,17].

The effort to synthesize the coaching literature in healthcare has also been subject to the problems of unclear naming and various methods of delivery. Prior to the effort to propose a definition of HWC by Wolever et al., reviews typically narrowed their inclusion by intervention name. For example, Ammentorp et al. conducted a review of “life coaching.” [18]. Kivela et al., examined “health coaching” in their systematic review [19]. Additionally, previous reviews have excluded studies using peer- or lay-coaches, yet many current interventions are using lay coaches [20,21]. Finally, while some reviews have meta-analyzed condition specific outcomes (e.g. in diabetes [22,23], recent reviews to summarize HWC have not meta-analyzed patient-important outcomes [19,24]. Previous research on patient-important outcomes has defined them as: “a characteristic or variable that reflects how a patient feels, functions, or survives, that is, all outcomes leading to important changes for patient life.” [25]. Additionally, past reviews have not used theories pertaining to multimorbidity, such as the TPC, to elucidate HWC components that may influence the efficacy of the intervention for the population of people living with MCC.

Therefore, we conducted a systematic review and meta-analysis seeking to summarize for patients with chronic conditions, the impact of HWC on patient important outcomes including quality of life (QoL), self-efficacy (SE), anxiety, depression compared to any intervention without coaching (e.g., patient education). We sought to include all HWC interventions delivered in alignment with the Wolever et al., 2013 conceptual definition [8], delivered by professional or peer coaches. Finally, we aimed to explore what, if any attention, to constructs in the TPC were associated with efficacious HWC interventions.

2. Methods

This systematic review adheres to Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [26], and it was guided by a detailed protocol that was published prior to the conduct of this review [27]. (PROSPERO registration: CRD42016039730).

2.1. Study eligibility

2.1.1. Types of studies

Eligible studies were randomized controlled trials that compared HWC to standard care or any interventions without coaching. We excluded non-English publications. No studies were excluded based on the risk of bias assessment.

2.1.2. Types of participants

We included studies that used HWC as an intervention for adults (18 +) with one or more chronic conditions. We used the Agency for Healthcare Quality and Research’s definition of a chronic condition as one “lasting 12 months or more and limiting self-care, independent living, or requires ongoing medical intervention.” [28]. We excluded studies that were delivered in a healthcare setting but were not delivered in the context of treatment for a chronic condition (e.g., general health maintenance, stress management, etc.).

2.1.3. Types of interventions

We included studies that used HWC as an intervention for individuals with chronic conditions. Studies were included regardless of delivery method (e.g., telephone vs. in-person) or format (1:1 format, group format, or a combination). In order to be included, studies had to (1) be patient-centered in that patients must at least partially determine their goals; (2) use self-discovery or active learning processes together with content education to work toward patient goals; (3) have a component of behavioral self-monitoring to increase accountability; and (4) pursue these activities in the context of an interpersonal relationship with a coach [8]. We included studies regardless of the training of the coach delivering the intervention, which included health professionals, certified coaches, community health workers, peer-coaches, and others.

2.1.4. Types of comparators

We included studies that compared HWC to standard clinical care or another intervention without coaching (e.g., patient education, counseling, chronic disease self-management).

2.1.5. Types of outcome measures

QoL, self-efficacy (individuals’ belief in his or her capacity to execute behaviors), depression and anxiety symptoms evaluated by any self-reported scale at any time points.

2.2. Search strategy

Our search strategy was developed by an expert reference librarian (LP) in collaboration with the investigators of the study and included a mix of keywords and controlled vocabulary. A complete search strategy is included in Appendix 1. We searched the following databases from January 2005-April 2023: Ovid MEDLINE, Ovid EMBASE, CINAHL, Ovid PsycINFO, Ovid Cochrane CENTRAL, and Scopus.

2.3. Selection of studies

After the studies were exported from the databases, they were imported into reference management software (Endnote), at which time they were de-duplicated using the native de-duplication function within the software followed by a manual review. The remaining studies were then imported into a systematic review software (DistillerSR, Ottawa, Canada). We screened studies in two phases: abstract and full text. Reviewers (KB, SB, SA, NA, HD, JM, FR, GE, HV) were first trained regarding study purpose, inclusion, and exclusion criteria, and then screened a set of abstracts in triplicate to ensure good inter-rater reliability. Abstracts were then screened independently and in duplicate by the reviewers; abstracts, where reviewers agreed, were moved to full-text screening, or excluded. Abstracts with disagreements between reviewers were automatically moved to the full-text screening phase. Each full-text article was screened independently and in duplicate. Full-text articles where the two reviewers were not in agreement were discussed with a third researcher that was invited to review disagreements.

2.4. Data extraction

Data were extracted using the same systematic review software (DistillerSR, Ottawa, Canada). Data extraction was undertaken independently by reviewers in pairs (KB, SB, GS, SA, MA, AA, NA, HD, JM, FR, GE, HV); discrepancies that could not be resolved via discussion were referred to a third reviewer. The data extraction template was piloted on several studies and modified as needed. We extracted lead author names, date, country in which the study was conducted, chronic condition(s) targeted, study aim, study design, sample subjects characteristics including total number, age, sex, loss to follow-up, ethnicity/race, sample income information, intervention and control characteristics including the person delivering the intervention, number of sessions prescribed, duration of session prescribed, mean number of sessions, mean duration of sessions, length of intervention, and control group characteristics. We extracted all measures of QoL, self-efficacy, depression, and anxiety, as well as the instrument used. Two reviewers (KB and MG) evaluated intervention descriptions for meeting TPC constructs. Disagreements regarding intervention components and theoretical constructs were resolved by consensus between the two reviewers.

2.5. Risk of bias assessment

We used the Cochrane Collaboration’s tool for assessing the risk of bias (RoB 2), which prompts assessment of: bias arising from the randomization process, deviation from intended interventions, missing outcome data, measurement of the outcome(s), and selection of the reported result [29].

2.6. Statistical analysis

To evaluate the effectiveness of the interventions, we calculated standardized mean difference (SMD) using the Cohen’s d method [30]. Appropriate transformations were made to ensure that the direction of each scale was the same (mean scores multiplied by −1).

The DerSimonian-Laird random effect model was used to pool SMD from the included studies [31]. We considered the standardized effect to be small, moderate, and large using cutoffs of 0.2, 0.5 and 0.8; respectively [32].

Heterogeneity, I² was also calculated to evaluate the percentage of variability in the effect estimate that is due to heterogeneity rather than chance, in which > 50 % suggests substantial heterogeneity [33]. All statistical analyses were conducted using RevMan (version 5.4; The Cochrane Collaboration, 2020) and the meta package in R (version 3.4.3; R Project for Statistical Computing).

The insufficient number of studies available for each time point hindered the feasibility of conducting a meta-regression encompassing all five capacity constructs. Therefore, after statistical analyses were conducted for each outcome, we narratively summarized the constructs that were present in studies with demonstrated effect on the outcome of interest at one or more timepoints.

2.7. Certainty of evidence

We evaluated the certainty of evidence using the GRADE for complex interventions [34]. Certainty of evidence from randomized trials starts at high and can be rated down for methodological limitations, imprecision, indirectness, inconsistency, or publication bias [35].

3. Results

The search strategy yielded 15,034 papers, and two more papers were found through a snowball strategy, of which 10,139 remained after removing duplicates. After the abstract screening, 1337 studies remained, of which 30 met the inclusion criteria and reported an outcome of interest. Three studies were excluded from the meta-analysis on the basis that they presented change scores as opposed to final means and standard deviations [36–38] and one study was excluded because the questionnaire used did not provide an overall score so pooling wasn’t possible [39]. The PRISMA flow diagram is depicted in Appendix 2.

Studies were published between 2006 and 2022: thirteen studies in the United States [36–38,40–49], five studies in Australia [39,50–53], five studies in the Netherlands [54–58], three studies were conducted in the United Kingdom [59–61], one study in Turkey [62], one study in Germany [63], one study in Nigeria [64], and one study in South Korea [65]. The total sample size of the 30 included studies was 8662 with study sample sizes ranging from 26 to 2707 (median = 148.5). Table 1 shows the characteristics and main findings of the included studies.

Table 1.

Main characteristics and findings of the included studies.

Author, year	Country	Chronic condition	Total N, Mean age, %male	Intervention group	Control group	Aim of the study
Bennell et al. [50]	Australia	Knee Osteoarthritis	N = 168, 62.2 years, 36.9 %	Physiotherapy group with telephone coaching: 5 individual sessions, and additionally received 6 telephone- delivered coaching sessions over 6 months (approximately weeks 2, 4, 8, 13, 21 and 25). The mean attended sessions was 4.3 ± 1.4, and the mean number of telephone sessions was 5.4 ± 2 with a mean call duration of 24.1 min. The telephone calls were made by 5 female coaches (3 nurses, 1 occupational therapist, 1 health psychologist)	Physiotherapy group: Visited a project physiotherapist for 5 individual 30-minute sessions over 6 months (weeks 1, 3, 7, 12 and 20). The mean attended sessions was 4.4 ± 1.2.	Evaluate if simultaneous telephone coaching improves the clinical effectiveness of physiotherapist for knee osteoarthritis
Benzo et al. [41]	United States	Chronic Obstructive Pulmonary Disease	N = 215 68 years, 45.11 %	Health coaching: One initial visit face to face, and subsequently 21 calls by the coach, a nurse, or a respiratory therapist. Also, they received care accordance with the Global Initiative for Chronic Obstructive Lung Disease and revered for a conventional pulmonary rehabilitation. They established as a complete intervention 15 of 21 calls, which only 85 % of the participants reached.	Standard of care: The group received care accordance with the Global Initiative for Chronic Obstructive Lung Disease and revered for a conventional pulmonary rehabilitation.	To determine the effect of comprehensive health coaching on the rate of COPD readmissions.
Benzo et al. [36]	United States	Chronic Obstructive Pulmonary Disease	N = 146, 68.94 years, 48.63 %	Home-Based system: Participants in the intervention arm received an Android tablet with an interactive application that presented a daily to-do-list with complete daily exercises. All activities and vital signs were visible to the participants’ health coaches. Health coaching was done weekly over the telephone to review patient activity and symptoms and to allow an opportunity for the participants to define their goal. Program’s duration was 8 weeks.	Control group: Participants in this group had an 8-week control period and was compassionately offered the intervention after the control period.	To assess feasibility, uptake, adherence, and patient experience alongside the effectiveness of a home-based rehabilitation program with health coaching.
Benzo et al. [40]	United States	Chronic Obstructive Pulmonary Disease	N = 375, 69 years, 43.46 %	Intervention group: Participants had a weekly health coach call aimed to be a behavior change intervention, including a reflection of daily steps, symptoms logged daily in a table, compliance and feedback to the flexibility and balance video-guided exercises and setting weekly goals.	Control group: Individuals had usual care and received an educational packet of 12 self-management themes for weekly self-study to match the intervention.	To test the effect of a home-based program of RPM with HC on the physical and emotional quality of life in individuals with stable COPD.
Blackberry et al. [39]	Australia	Type 2 diabetes	N = 473, 62.8 years, 57 %	Telephone coaching: This group was led by coaching nurses through telephone calls, being the median number of total coaching sessions was 3 (IQR, 1–5). The time spent by practice nurses in delivering coaching ranged between 10 and 120 min per session, with a median of 30	Usual general practice care: Usual general practice care was provided, referral to diabetes educators, dietitians, and diabetes specialists as part of the standard diabetes care provided by that practice.	To evaluate the effectiveness of goal focused telephone coaching by practice nurses in improving glycemic control in patients with T2DM.
Burgess et al. [42]	United States	Chronic Musculoskeletal Pain	N = 380, 58 years, 72 %	Intervention arm: Participants received a pedometer and tailored materials. Participants were coached to create and write action plans for their proposed walking activity and using templates contained in their workbooks. Coaches created plans to overcome barriers, strengthen helpful factors and involve friends and family members. The intervention consisted of 6, individual 30–60-minute telephone coaching sessions over a 8–14 week period.	Usual care: Participants received a brochure with advice and information about the benefits of walking and a pedometer to record their steps with.	To test whether a walking-focused, proactive coaching intervention based on a biopsychosocial framework, aimed at addressing contributors to racial disparities in pain, would, lead to greater improvement in chronic pain outcomes among Black patients, relative to usual care.
Cameron-Tucker et al. [51]	Australia	Chronic Obstructive Pulmonary Disease	N = 65, 69 years, 45 %	Tele-rehab: Participants were contacted via telephone by specifically trained community over the next 8–12 weeks, to support the home-walking action plan. A schedule of two calls weekly was suggested, with a minimum of four calls mutually agreed. There was a total of 231 telephone contacts, with a mean of 7 ± 3 calls per individual, each lasting a duration of 17 ± 9 min per call and 7 ± 6 days between calls.	Usual care: Participants in usual care group waited for 8–12 weeks prior to their scheduled pulmonary rehabilitation appointment without any additional contact	To evaluate the effectiveness of telephone health-mentoring targeting home-based walking compared to usual waiting time in physical capacity.
Cinar et al. [62]	Turkey	Type 2 diabetes	N = 186, N/R N/R	Health Coaching Group: One professional health coach intervention for the participants. Participants had a face-to-face session with the coach 2 weeks after the baseline visit. Cinar et al., did no report numbers of visits, nor duration.	Health Education Group: Participants received standard lifestyle advice referring to oral health care practices, diet, and physical exercise. All participants in both groups were supported by 8 text messages at the initiation phases.	To assess how overall patient satisfaction with medical provider- (dentist- and physician-) patient communication can affect oral health, diabetes, and psychobehavioral measures among T2DM patients.
Hawkes et al. [52]	Australia	Colorectal cancer	N = 410, 66.3 years, 54 %	Health Coaching: Participants received 11 telephone-delivered health coach sessions over a 6-month period. All the health coaches had university degrees in nursing, psychology, or health promotion.	Usual care: Participants received four freely available educational brochures produced by Cancer Council Australia on understanding CRC and cutting cancer risk, diet, and PA.	To determine the effects of a telephone-delivered multiple health behavior change intervention on health and behavioral outcomes among colorectal cancer survivors.
Houlihan et al. [37]	United States	Chronic Spinal Cord Injury	N = 84, 45.6 years, 26.2 %	My Care my Call: Participants completed 12 ± 5 calls, averaging 21.8 min each over 6 months. The calls were made by peer health coaches.	Control group: Participants received the usual care with no new peer or other support services.	The aim was to evaluate if the intervention had a positive impact on physical activation in adults with CSCI.
Humalda et al. [54]	The Netherlands	Chronic kidney disease	N = 99, 56.6 years, 84.04 %	E-coaching: Coaching was done by dietitians, lifestyle coaches, or research nurses; and participants in the intervention group received 2.8 ± 1.2 sessions of e-coaching. Additionally, they received 2 group coaching sessions of 2 h.	Routine care: Humalda et al., did not define it.	To evaluate a self-management approach for dietary sodium restriction in patients with CKD.
Johnson-Warrington et al. [59]	United Kingdom	Chronic Obstructive Pulmonary Disease	N = 78, 68 years, 35.8 %	SPACE for COPD: Participants were introduced to the manual and exercises by a trained physiotherapist in a one-to-one session lasting 30–45 min. Additionally, they received structured phone calls within 72 h and at 2 weeks, 4 weeks, 6 weeks, 8 weeks, and 10 weeks from hospital discharge. Johnson-Warrington et al., did not specify the mean of attendance, nor the way of these sessions.	Usual care: This consisted of a follow-up appointment with the community COPD team or telephone follow-up after an inpatient review by a respiratory nurse specialist and an outpatient consultant review.	To investigate if SPACE for COPD employed upon hospital discharge would reduce readmission rates at 3 months.
Jolly et al. [60]	United Kingdom	Chronic Obstructive Pulmonary Disease	N = 577, 70.4 years, 63.5%	Intervention group: The intervention consisted of telephone health coaching delivered by a nurse with supporting written documents, a pedometer, and a self-monitoring diary. The first telephone coaching session at one week after randomization aimed to last 35–60 min, followed by a 15–20-minute telephone session at weeks 3, 7, and 11. 86.4 % of the scheduled calls were delivered and 75.4 % of participants received all four calls. The average duration of calls was 39.2 ± 10.7 min for the first call, then 23.8 ± 9.2, 21.4 ± 8.6, and 20.6 ± 8.7 min for the second, third, and final calls, respectively.	Usual care: The usual care group received a standard information leaflet about self-management of COPD. The leaflet gave a definition of COPD, a detailed description of associated symptoms, how the illness can be managed with the use of inhalers, how to treat exacerbations, and details of other resources.	To evaluate the effectiveness of telephone health coaching delivered by a nurse to support self-management in a primary care population with mild symptoms of chronic obstructive pulmonary disease.
Kempf et al. [63]	Germany	Type 2 Diabetes	N = 202, 59.5 years, 45 %	Telemedical Lifestyle intervention Program (TeLiPro): The TeLiPro group received the same as control group and additionally received a blood glucose meter. Also, they received dietary intervention to achieve an initial weight reduction and weekly care calls from trained diabetes coaches. The mean duration of care calls in this group was 17 min (range 12–30).	Routine care: Participants in control group received a self-management guide, a weighing scale, and a step counter.	To evaluate the efficacy of the Telemedical Lifestyle intervention Program in improving metabolic control in advanced-stage type 2 diabetes.
Knittle et al. [55]	The Netherlands	Rheumatoid Arthritis	N = 78, 62.7 years, 33.33 %	Coaching group: The intervention in the treatment group consisted of 7 sessions. Those sessions were about education, motivational interview, self-regulation coaching sessions, and telephone calls, which were delivered by a physical therapist and a rheumatology nurse.	Control group: Patients in control group at the beginning attended a small group educational session delivered by a physical therapist.	To evaluate the effects of targeting both the motivation and action phases of behavior change in a 5-week intervention to increase physical activity among patients with rheumatoid arthritis not meeting current physical activity recommendations.
Mayberry et al. [38]	United States	Type 2 Diabetes	N = 506, 56.93 years, 45.84 %	FAMS intervention: Included monthly family/friend-focused phone coaching sessions, text messages tailored assess the goal set during coaching and an option to invite an adult support person to receive text messages. Coaches had the same components: a check in to assess participants baseline status; goal setting; discuss family/friends’ involvement; skill building and a verbal contract for commitment. Coaching occurred monthly for six full sessions, each lasting 20–30 mins.	Rapid Encouragement/Education and Communications for Health: Provided daily text message support with content to address barriers for self-care behaviors. It also sent daily text messages asking if the participant took their medications; followed by weekly feedback. Control group: Participants received treatment as usual.	To determine whether FAMS was acceptable and well received in a diverse sample of adults with Type 2 diabetes, to ensure that the intervention improved the response/actions from family/ friends as patients meet self-care goals, and to ascertain if improvements in family/friend involvement mediated improvements in dietary behavior, physical activity, and diabetes selfefficacy over 6 months.
McGonagle et al. [43]	United States	Any Chronic Illness	N = 59, 38.6 years, 14 %	Telephone Coaching: The intervention included six 1-hr phone-based coaching sessions over 12 weeks (one session every other week). The coach was certified by the International Coach Federation and followed their competency guidelines. McGonagle et al., did not report the adherence, nor the mean duration of sessions.	Control group: The waitlisted control group participants waited for 12 weeks. They did not receive any interaction, just another survey after the 12 weeks of intervention in the other group.	To evaluate a coaching intervention designed to help individual workers with chronic illnesses manage challenges stemming from working with illness.
Meuleman et al. [56]	The Netherlands	Chronic kidney disease	N = 151, 55.1 years, 81.9 %	Coaching group: Patients in the intervention group received the consultations with the nephrologist every 3–6 months, as the control group, and 3-month self-management intervention. During the intervention, patients were coupled with one of 4 personal coaches: 3 health psychologists and 1 dietician, all trained in motivational interviewing techniques. At the end of the study, only 82.1 % received the intervention according to protocol.	Control group: Regular care consisted of consultations with the nephrologist every 3–6 months and, if necessary, nutrition counseling by a dietician.	To evaluate the sodium excretion and blood pressure after the 3-month intervention and at 6-month follow-up.
Nguyen et al. [44]	United States	Chronic Obstructive Pulmonary Disease	N = 2707, 72 years, 46.3 %	Physical Activity Coaching (Walk On): Patients in intervention group received 4 weekly coaching phone calls during 1–5 weeks, and just 83 % achieved that. Between weeks 6–52 participants received a median (IQR) of 12 (7–20) phone calls. Additionally, 42 % of the participants attended at least one monthly group visits.	Standard of care: The standard care patients received their routine care and had access to all health services in accordance with their health plan.	To evaluate the long-term effectiveness of a community-based PA coaching intervention in patients with COPD.
Nishita et al. [45]	United States	Diabetes	N = 190, 48.5 years, 37 %	Coaching group: Participants determined the frequency of interactions with both the life coach and pharmacist, depending on their individual need. Life coaches’ mean response was 3.25 ± 0.46 and pharmacists’ mean response was 3 ± 0 reflected good adherence to the intended program model.	Control group: Participants at control group did not receive any intervention, they used their current control diabetes methods.	To examine the influence of life coaching and pharmacist counseling on the physical and psychosocial health of employed diabetics.
Olivier et al. [46]	United States	Coronary Artery Disease	N = 26, 65.65 years, 80.76 %	Intervention group: Participants received virtual health care in addition to usual care. A mobile application that included components of home-based cardiac rehabilitation that were protocoled in the program coaching guide was installed in the participants mobile. Participants were also paired to a cardiac rehabilitation coach, which offered a structured coaching intervention. An action plan was jointly created via telephone interview and participants received recommendations for physical activity via the app.	Control group: Participants received usual care, determined by the treating physician.	To evaluate the effectiveness of an intervention with a smartphone- based heart disease therapeutic for virtual health coaching, compared with physician- or nurse-guided standard of care, to improve cardiovascular medication adherence and cardiovascular risk factors in participants with coronary artery disease.
Onyechi et al. [64]	Nigeria	Type 2 Diabetes	N = 80, 52.7 years, 31.25 %	Cognitive behavioral coaching program: The participants in the treatment group took part in the cognitive behavioral coaching program for 18 sessions (sessions were held twice per week) for 9 consecutive weeks that lasted 50 min each. Nwamaka et al., did not specify the mean of attendance, nor the way of these sessions.	Conventional counseling approach: The researchers gave advice and instruction about managing depression to guide their judgment to make autonomous decisions about dealing with depression. Participants took part in a conventional counseling for 18 sessions (sessions were held twice per week) for 9 consecutive weeks that lasted 50 min each.	To investigate the effects of cognitive behavioral coaching program on depressive symptoms in a sample of the Type 2 diabetic inpatients
Park et al. [65]	South Korea	Multimorbidity	N = 50, 77.6 years, 20.9 %	Health coaching self-management program (HCSMP): Each participant received eight face-to-face coaching sessions over the intervention period, which were led by geriatric nurses’ specialist. The mean number of sessions attended was 7.4 ± 2 %, and 44 % participants attended more than 90 % of their sessions.	Conventional group: Participants in the conventional group were asked to maintain their regular lifestyle including dietary and exercise habits for 8 weeks until they were reexamined.	To examine the effectiveness of a health coaching self-management program for older adults with multimorbidity in nursing homes.
Ream et al. [61]	United Kingdom	Cancer	N = 103, 64.1 years, 55 %	Intervention group: Monthly domiciliary meetings which facilitated provision of advice and coaching in strategies to manage fatigue by supported nurses. Ream et al., did not establish the mean nor duration of those meetings.	Control group: Ream et al., did not describe the intervention or approach of patients in control group.	To evaluate if a supportive intervention reduces the symptom of fatigue in patients undergoing chemotherapy.
Sacco et al. [47]	United States	Type 2 diabetes	N = 62, 52 years, 42 %	Coaching group: Patients received one phone call per week for the first 3 months, and one bi-weekly call for the remaining 3 months by undergraduates in psychology and a diabetes coach. The mean time per session was 17.38 ± 2.94 min, the average number of coaching sessions was 16.10 ± 2.28, and the mean duration of the coaching intervention was 24.3 ± 2.3 weeks.	Usual care: The control group received treatment as usual from a board- certified endocrinologist and completed the same pre- and post-test measures, approximately 6 months apart.	To evaluate the effects of a brief, regular, proactive, telephone “coaching” intervention on diabetes adherence, glycemic control, diabetes-related medical symptoms, and depressive symptoms.
Swoboda et al. [48]	United States	Type 2 Diabetes	N = 60, 56 years, 28.33 %	Single goal (SG) intervention: Participants received one baseline in-person goal- setting and decision coaching session to encourage lifestyle change followed by seven biweekly coaching calls delivered by a registered dietitian. Swoboda et al., did not define the mean nor duration of those meetings.	Attention control group: Participants received a guide to local health care resources and completed interviews that focused on discussion of community and public health resources.	Evaluate a 16-week decision support and goal-setting intervention to compare diet quality, decision, and diabetes- related outcomes to a control group.
Van der Wulp [57]	The Netherlands	Type 2 Diabetes	N = 133, 61 years, 54.6 %	Self-management coaching: Patients in intervention group received three training sessions, each lasting 3.5 hr by an experienced peer (expert patient). Van der Wulp et al., did not define the mean, duration of those meetings, nor the capacities of and expert patients.	Control group: Participants allocated to the control group received the same medical care as participants from the intervention group.	To study the effectiveness of a peer-led self-management coaching intervention in recently diagnosed patients with Type 2 diabetes.
Van Nimwegen et al. [58]	The Netherlands	Parkinson’s disease	N = 586, 65.5 years, 65 %	ParkFit programme: Activity coaches guided each patient towards a more active lifestyle during monthly personal coaching sessions, also it included regular physiotherapy sessions. The mean number of annual individual visits to the physiotherapist was 13.6	Control intervention: It consisted of a general physiotherapy program aimed at promoting safety of movements, according to the evidence-based guideline.	To evaluate whether a multifaceted behavioral change program increases physical activities in patients with Parkinson’s disease.
Varney et al. [53]	Australia	Type 2 diabetes	N = 94, 64.1 years, 68 %	Coaching intervention: Intervention group participants received 6 months of telephone coaching, delivered by a dietitian with experience in cardiovascular disease and T2DM. Initial coaching sessions occurred within 2 weeks of randomization and subsequent sessions occurred approximately monthly. Participants received 6.0 (range 4–9) coaching sessions.	Control group: Control group participants could access STV usual care services, including a diabetes clinic staffed by endocrinologists, diabetes educators and dietitians. They did not receive the telephone coaching intervention, or any contact from the researchers, except for telephone calls to arrange baseline, 6- and 12-month assessment appointments.	To measure the effect of a 6-month telephone coaching intervention on glycemic control, risk factor status and adherence to diabetes management.
Young et al. [49]	United States	Type 2 Diabetes	N = 319, 59.01 years, 52.7 %	Intervention group: The intervention included a nurse health coaching and a mHealth technology to track PGHD and integrate these data into the EHR. The participants had an in-person orientation with the nurse coach, followed by telephone sessions every 2 weeks for 3 months. Across coaching sessions, participants averaged 172 min of nurse coaching.	Usual care: Participants in this group received usual care through their primary care clinic, which comprised standard health care visits with providers and access to classes, resources, and services.	To evaluate the effectiveness of a nurse coaching program using motivational interviewing paired with mobile health (mHealth) technology on diabetes selfefficacy and self-management for persons with type 2 diabetes.

± values: mean ± standard deviation. Abbreviations: COPD, Chronic Obstructive Pulmonary Disease; T2DM, Type 2 diabetes mellitus; IQR, Inter Quartile Range; CRC, Colorectal cancer; CKD, Chronic Kidney Disease; CSCI, Chronic spinal cord injury; PA, physical activity; PGHD, patient-generated health data; EHR, electronic health record; FAMS, Family-focused Add-on to Motivate Self-care; PEERS, Program for the Education and Enrichment of Relational Skills; ASD, autism spectrum disorder. P value reported for the main outcome of the study. N reported for the total randomized participants.

Risk of bias assessment and evidence certainty assessment are illustrated in Appendices 3 and 4. The certainty in the evidence for most outcomes was either very low or low primarily due to the high risk of bias, heterogeneity, and imprecision.

3.1. Intervention Effect on Quality of Life (QoL)

Fourteen studies reported QoL outcomes [39–41,44,45,50–52,54,56, 58,60,62,63]. HWC was associated with a moderate improvement in QoL scores at 3 months (n = 5, SMD 0.62, 95 % CI 0.22–1.02, p = 0.002, I² = 86.3 %), but not at 6, 12 and 18 months (Fig. 1).

Fig. 1.

SMD and 95 % CI for effect of HWC on Quality of life.

3.2. Intervention effect on self-efficacy

Thirteen studies reported self-efficacy outcomes [39,43,45,47–49, 55–57,59,60,62,65]. HWC was associated with a small improvement in patient-reported self-efficacy at 1.5 months (n = 2, SMD 0.38, 95 % CI 0.03–0.73, p = 0.033, I² = 0 %). No significant effects were found on longer-term patient-reported self-efficacy. (Fig. 2).

Fig. 2.

SMD and 95 % CI for effect of HWC on Self-efficacy.

3.3. Intervention Effect on Depressive symptoms

Fifteen studies reported outcomes of patient’s depressive symptoms [39,40,42,44,47–49,53,55,57,59–61,64,65]. HWC was associated with a moderate improvement in depressive symptoms at 3 and 6 months (n = 8, SMD 0.67 95 % CI 0.13–1.20, p = 0.01, I2 = 95 %) and (n = 8, SMD 0.72, 95 % CI 0.19–1.24, p = 0.006, I2 = 95.6 %), respectively. Furthermore, HWC was associated with a small improvement in depressive symptoms at 12 months (n = 4, SMD 0.41, 95 % CI 0.09–0.73, p = 0.01, I2 = 63.8 %), but not at 1.5, and 18 months (Fig. 3).

Fig. 3.

SMD and 95 % CI f or effect of HWC on patient’s depressive symptoms].

3.4. Intervention effect on anxiety symptoms

Seven studies reported anxiety outcomes.[40,42,44,49,59–61] It was possible to perform an analysis of the effect of HWC on patients’ anxiety symptoms at 3 and 12 months, no statistically significant effects were found. A greater score indicates decreased anxiety symptoms. (Fig. 4).

Fig. 4.

SMD and 95 % CI for effect of HWC on patient’s anxiety symptoms.

3.5. Capacity constructs and their relationship with outcomes

Appendix 5 shows each included study and the constructs of the TPC that were acted upon in their described intervention. All interventions acted upon the construct of “Work” as this construct is aligned with the Wolever et al. definition of coaching, which was a requirement for inclusion in the review. Most studies included attention to at least one additional capacity construct. Most commonly HWC interventions included attention to the Resources construct (25 out of 30). Other capacity constructs were included in interventions less commonly. Biography was attended to in eight studies, [40,41,43,45,52,61,64,65] Environment in five studies [39,41,49,53,65], and Social in eight [38, 41,42,44,54–56,65]. Of the 21 significant outcomes, 14 occurred in studies with 3 or more of the BREWS constructs represented in their interventions (67 %).

Exploring the characteristics of the 14 studies that evaluated QoL, we observed that 6 out of 14 studies found improvement in favor of the coaching intervention at one or more timepoints [40,41,54,56,62,63]. All included attention to Work and Resources; three of the six included additional attention to the Social construct [41,54,56]. One of the two studies that found improved QoL at longer term follow-up, included attention to all five capacity constructs [41].

Six out of 14 studies that evaluated self-efficacy found improvement at one or more timepoints after HWC [45,47,49,55,56,62]. The majority (83 %) included attention to the Work and Resources constructs. However, most also included at least attention to one other construct including Environment [49], Biography [45], or Social [55,56].

Five out of 15 studies found improvement in favor of the coaching intervention when the outcome was depression [40,53,57,60,64]. One study included only the Work construct[57] two included two constructs [53,60], and two included three constructs [40,64]. None of the studies found improvement in anxiety outcomes; therefore, exploration of capacity constructs was not possible for this outcome.

4. Discussion and conclusion

4.1. Discussion

This systematic review and meta-analysis of HWC interventions found that HWC interventions significantly improved patients’ QoL at 3 months, self-efficacy at 1.5 months, and depression symptoms at 3, 6, and 12 but not at longer follow up periods. However, few studies included long-term follow up. HWC did not significantly impact anxiety at any time. While all studies included attention to the Work construct in the TPC by virtue of the inclusion criteria, the majority also included some attention to the Resources construct. In studies demonstrating significant improvement across one or more outcome measures at any time point, the majority included attention to at least three capacity constructs.

This review includes numerous limitations that should be considered in interpreting its findings. First, we included studies that measured self-reported patient outcomes using a variety of scales. We made the assumption that all scales truly captured the same underlying construct. Second, we included all eligible studies regardless of whether they had a high or low risk of bias. Third, given our aim to be as comprehensive as possible, studies across a wide variety of populations were included. However, because of the small number of studies included overall, we were unable to explore the effects within population subgroups.

Fourth, two studies are noticeably different than others when examining their SMD. In the study conducted by Meuleman, et al. to examine coaching’s impact on sodium intake for patients with chronic disease, there is a notable positive effect on quality of life at 3 months, yet at 6 months there is a small negative effect [56]. In this study, participant QoL was similar at baseline between groups [56]. In the coaching group, mean QoL improved at 3 months while in the intervention group it declined [56]. At 6 month follow up, the coaching group’s gains in mean QoL were depleted and lower than both their baseline scores and 6 month mean QoL in the control group [56]. While it is unclear what aspects of this intervention or population are driving this change, a similar pattern is notable in the study’s primary outcome, sodium excretion rate, as well [56].

The study conducted by Onyechi et al. examined the effect coaching with patients with Type 2 diabetes who were both severely depressed and impatient at the time of enrollment [64]. At baseline, depression symptoms were similar across groups [64]. However, at 3 and 6 month follow ups, patients in the coaching group had marked decreases in depression compared to the control group which had stable depression scores over time [64]. One possibility for these differences is the unique population enrolled. Almost all other included studies were conducted in the outpatient setting and their inclusion criteria did not include severe depression. Therefore, the high levels of depression at baseline in this single study may have produced a greater reduction in depressive symptoms in the intervention group compared to the effects demonstrated in other studies that began with a more heterogeneous distribution of depression symptoms.

Finally, and perhaps the greatest limitation of this review is that it relies on reported characteristics within studies to assess which aspects of patient capacity were attended to in the HWC intervention. Until recently, the reporting of intervention characteristics was not standardized, and reporting is often insufficient to assess intervention characteristics. Despite these key limitations, this review also has numerous strengths compared to previous reviews. Specifically, we sought to include all chronic conditions, rather than isolate our review to a single condition. Further we used the first standardized definition of HWC for study inclusion, and we meta-analyzed four patient-important outcomes.

This is the first systematic review to meta-analyze coaching’s impact on patient-important outcomes across chronic condition types. Most similar to ours is the 2014 Kivela et al. review, which reported that 11 of 13 studies showed significant improvement in patients’ physiological, behavioral, psychological, and social outcomes, but no meta-analyses were conducted [19]. Furthermore, Kivela’s review excluded studies of patients with mental illness or disability and studies using peer-coaches [19], whereas our review included both. Our results echo their conclusions that HWC has short-term benefits when treating chronic illness. Additionally, a recent HWC compendium also noted mostly positive outcomes in the psychological domains (e.g., depression, QoL), amongst multiple disease populations, including cancer and diabetes, without meta-analyses [24]. Therefore, our research strengthens previous conclusions regarding HWC’s potential impact on chronic illness populations. However, our review also highlighted limitations in the current evidence base for HWC, including imprecision and heterogeneity. These limitations should give pause to those implementing HWC in clinical practice to ensure that their implementation is in alignment with HWC best practices which have been standardized [6] Additionally, these limitations highlight aspects of future studies that can strengthened in future research conducted to test HWC.

4.2. Conclusion

HWC is an intervention with the potential to benefit patients living with chronic illness. Our systematic review and meta-analysis sought to review the evidence regarding the effectiveness of HWC in a diverse chronic illness population on depression, anxiety, self-efficacy, and QoL. We found that HWC is effective on short-term patient outcomes of self-efficacy, depression, and quality of life. HWC did not impact patient anxiety at any time point. We did not find evidence of long-term benefit, but this finding may be due to the paucity of studies examining longer-term outcomes. We caution that the certainty in the evidence for most outcomes was either very low or low primarily due to the high risk of bias, heterogeneity, and imprecision. In exploring the relationship between attention to constructs in the TPC and patient outcomes, we found that most commonly, effective interventions included components related to at least three capacity constructs. Future HWC research should seek to standardize the reporting of interventions and outcome measurement, extend follow-up times, and employ comparative effectiveness designs to test the varied impact of intervention components that address different capacity constructs.

4.3. Practice Implications

The body of literature on HWC for chronic conditions is heterogeneous. This heterogeneity is, in part, due to a lack of standardization of intervention delivery and reporting as well as outcome measures to assess its effect. To advance our understanding of HWC efforts should be made to standardize the definition of HWC as well as the measures used in trials of HWC; efforts like the COMET initiative support the development of standardized outcome sets, yet none exist for HWC [66]. In our review, we found 11 scales were used to collect patient reported QoL alone, highlighting the severity of this problem. In addition to standardization, trialists conducting trials in HWC should ensure rigorous trial design including utilizing comparative effectiveness designs, having an adequate duration intervention and follow-up, and complete reporting of interventions. Intervention reporting could be guided by frameworks such as TIDieR [67]. Additionally, heterogeneity may be related to the varying nature of patient-chosen goals. All included studies allowed patients to partially or completely choose their goals; some may have been more suited to improvement and alignment with collected outcomes versus others.

Finally, interventions should be designed such that they are aligned with emerging theories relevant to patients with MCC such as the TPC [15]. The majority of interventions demonstrating efficacy included attention to at least three constructs included in the TPC, typically a combination of Work and Resources plus one amongst Biography, Environment, or Social. Few interventions include attention to all constructs. Future HWC trials could determine a priori which constructs will be attended to as part of the intervention(s). Further, a comparative effectiveness approach could be used to elucidate which constructs are most useful or if additional constructs have a cumulative impact on patient-important outcomes.

Few studies reported long-term follow-up. Given positive short-term outcomes on self-efficacy, depression and on QoL, it would be useful for future studies to follow patients for longer periods to determine if the effects of coaching persist long-term or if booster interventions are required. Further, because chronic disease is lifelong, the duration of the coaching intervention and the durability of its impact post-intervention is important for clinical decision-making. From the evidence today, we cannot yet conclude the appropriate length for HWC interventions or whether intervention boosters are required.