Behavioural Change Techniques in Health Coaching-Based Interventions for Type 2 Diabetes: A Systematic Review and Meta-Analysis

To identify the relevant literature, a series of systematic searches was conducted on PsycINFO, Medline (Ovid) and the Web of Science. The searches were conducted using the keywords and their combinations. Medline key search terms included: “type II diabetes mellitus,” “non-insulin dependent diabetes mellitus,” “Diabetes Mellitus, Type 2/ or diabetes,” “Coaching,” “Health Coaching,” and “personal coach*” (see Supplementary Material 1 for more details on Medline search strategy). A manual back chaining was utilised as an additional step to supplement the database searches find relevant literature. This involved examining the list of all the references in the included studies, including potential citations within each article and other relevant reviews.

The current review focused on people with T2DM (Population), who received health coaching (Intervention) and who were compared with a usual care or active control group (Comparison) on HbA1c levels (Outcome). Studies were only included if they were peer-reviewed RCTs, reported changes in HbA1c, published in English from January 1950 and April 2022, included participants aged 18 years or older and employed health coaching to influence T2DM. For the purpose of this review, health coaching was defined as using client-centred sessions in which the coach uses coaching skills and techniques to enable the client to engage and work toward their intended goals. The start date of searching was purposely selected to cover all coaching terms, such as health counselling, coaching, personal coaching, and health promotion in published studies from the emerging time of health coaching in the early 1950s. Articles were excluded if participants did not have a diagnosis of T2DM; were not subject to health coaching interventions; self-management was not the targeted behaviour; included other variations of diabetes, e.g., gestational diabetes or type 1 diabetes mellitus; and HbA1c was not reported as an outcome measure. This review therefore included interventions that investigated the effectiveness of using the health coaching approach as a tool to impact the self-management of T2DM. Only RCT studies were included to explore effectiveness of the interventions and minimize the risk of bias [33].

Search results were initially screened against the inclusion criteria at title and abstract level. Full texts of these articles were screened next. Screening was completed independently by two researchers (AA, HH). The first author extracted data from the included studies, and then the second author reviewed the data for verification. Conflicts resolved by discussion between two reviewers (AA, HH). An independent reviewer (PN) conducted an additional step to double-check the extracted data. Data were systematically extracted using a prespecified extraction form (see Table 1 and Supplementary Material 2). Related studies (e.g., published protocols) were reviewed to extract further information. Relevant study information from the included articles was reviewed and data extracted (e.g., design, the theory or model used, BCTs, intervention structure, target behaviours, and outcome parameters) by two reviewers (AA, HH). The RCTs included were coded as to theories and BCTs used in the interventions as well as reported effects on glycaemic control. RCTs were also coded according to the modes of delivery, length and duration of the health coaching sessions.

Effect sizes (Cohen’s ​d) [54] for the included interventions were calculated in line with recommended procedures for pretest-posttest-control group designs (i.e., RCTs with pre- and post- measures of the outcome variable) [55] which control for baseline differences in the outcome measure. In particular, baseline mean HbA1c values were subtracted from follow-up mean values for the intervention and control groups, separately, and these new values used to compute the effect size difference. Baseline standard deviations were used to estimate the pooled standard deviation to account for the fact that, if the intervention changes the outcome at follow-up, variation in outcome scores is likely to be greater in the intervention compared to control group. An Excel spreadsheet was created to calculate effect size differences following Morris’ (16) formula based on data reported in the papers. Where baseline scores were not reported, effect sizes were based on follow-up scores using software available at www.​psychometrica.​de. The effect sizes were calculated so that positive effect sizes indicated greater reductions in HbA1c in the intervention group compared to the control group. As per Cohen’s guidelines, the intervention has a small effect size when d ≥ 0.20, a medium effect size when d ≥ 0.50, and a large effect size when d ≥ 0.80. Effect sizes of d < 0.20 were considered to be trivial.

The BCT taxonomy [25] was applied to the included studies to identify the use of BCTs. Two independent researchers (AA, HH) coded the intervention content reported in the methods section (intervention description) of each paper against the BCT taxonomy version 1 (BCTTv1), to identify the BCTs used in the health coaching interventions [25]. The coders followed the BCTTv1 guidance, for example, if a BCT was unclear (present or absent), it was coded as absent, as per the BCTTv1 guidance [25]. Both coders used Microsoft Excel (version 16.66.1) to generate a list of identified BCTs across all included interventions. Several discussion meetings were held to discuss the BCTs identified and to resolve any disagreements regarding the coded BCTs until reaching an agreement. A third independent reviewer (PN) was involved to confirm consensus decisions.

Meta-Essentials version 1.5 [56] was used to compute the sample-weighted average effect (Hedges g₊) of the health coaching interventions on HbA1c scores. Cochrane’s Q was used to test whether the effect sizes were heterogeneous and the I² statistic was used to assess the proportion of the variance in the effect sizes explained by any heterogeneity. Moderator analyses were then conducted to identify variables that accounted for any variability in effect sizes. For categorical moderators (e.g., presence or absence of a BCT) average effect sizes were calculated for each level of the moderator. The difference between the effect sizes was assessed using the Q statistic. The significance of continuous moderators was tested using meta-regression (see Tables 2 & 3).

Publication bias was assessed through visual inspection of the funnel plot (i.e., lack of asymmetry in the distribution of the studies) and Egger’s regression.

The Cochrane collaboration tool was used to assess the quality of the included studies [57].. Each study was rated based on specific criteria related to the quality of its methods and reporting, selection, performance, detection, attrition, reporting, and other biases. The assessment of study quality was evaluated by three reviewers (AA,EG,SC). See Table 4 and Fig. 2 for further details.

The search results yielded 1163 titles and abstracts through Medline, PsycINFO and the Web of Science. There were 145 full-text studies checked for eligibility and a total of 20 RCTs met inclusion criteria (see Fig. 1) [32].

Meta-analysis of 20 effect sizes from 20 unique studies, with a total sample of 3222 participants, indicated that, on average, health coaching interventions for T2DM have a small but statistically significant (positive) effect on reducing HbA1c (g₊ = 0.29, 95% CI: 0.18 to 0.40). Visual inspection of the funnel plot suggested that there was no asymmetry in the distribution of the studies and no risk of publication bias. Egger’s regression was also non-significant (p = 0.730), indicating lack of publication bias.

The effect sizes (d) of interventions ranged from d = − 0.05 to d = 0.78. None of the interventions had a large effect size [44], and only three had a medium effect size (d = 0.71 to d = 0.78) [42, 45, 51, 53]. The remaining 17 interventions had small (d ≥ 0.20) [36, 38‐40, 43, 46‐49, 52] or trivial (d < 0.20) effect sizes [34, 35, 37, 41, 50]. Cochrane’s Q was statistically significant (Q = 36.68, p = .009) suggesting that the effect sizes were heterogeneous and the I² statistic indicated that a proportion of the variance in the effect sizes was explained by this heterogeneity (I² = 48.20%), which indicates a need for moderation analysis to identify variables that account for the variability.

Table 1 reports the characteristics of included studies for both interventions (health coaching), and control groups (usual care), including sample size, mean age of participants, intervention duration, personnel, and mode of delivery (e.g., face-to-face, telephone-based, web-based). The included studies comprised 20 RCTs published between 1950 and 2022. A total of 3222 participants were included in the 20 studies, of whom 1674 were randomised to receive coaching interventions and 1548 were allocated to control groups. The majority of studies (n = 10) were conducted in the US [34‐43], two were conducted in Taiwan [44, 45], and the rest were conducted once in different countries including Turkey [46], Canada [47], South Korea [48], Norway [49], Finland [50], Germany [51], Belgium [52], and Australia [53]. In the 17 studies that reported gender of participants, 53% of participants were female. The mean age of the recruited participants was 59.3 (SD = 6.2). Due to the inconsistent reporting of other demographic and socioeconomic characteristics, such as education, ethnicity and income status, across the 20 papers we were unable to report them here. The recruitment of participants was varied and drawn from different communities including ethnic community centres [36], community health centres [34, 48, 49], community advertisement [43, 47, 49, 51], primary care or hospital clinics [38, 41, 45, 46, 53] and databases [40, 44, 50, 52]. For clinical factors, including HbA1c, there were no discernible changes between the intervention and control groups at baseline. The mean HbA1c level across all studies at baseline was 8.42% (SD = 0.78). The reduction in HbA1c found to be clinical significant in eight studies [36, 40, 42‐44, 46, 47, 51] (decrease of ≥5 mmol/mol )[58].

Moderation analysis of the sample characteristics indicated that intervention effectiveness was not related to age (β = 0.19, p = 0.442) or gender (β = − 0.13, p = 0.603). Moderation analysis of the study characteristics indicated that only the type of primary outcome measure was significantly related to intervention effectiveness (Q = 4.20, p = 0.040), such that studies including HbA1c as the primary outcome (g₊ = 0.32, k = 16) were more effective than studies with other primary outcomes (g₊ = 0.10, k = 4).

Health coaching was delivered through various methods including exclusive telephone-based [34, 39, 43, 47, 52, 53], exclusive web or mobile-based remote patient monitoring/electronic assistance (ERPM/EA) systems [37] or in combinations of face-to-face and telephone-based [36, 38, 40, 42, 44‐46]; face-to-face and ERPM/E A[48] telephone-based and ERPM/EA [49‐51] or face-to-face, telephone-based and ERPM/EA [35, 41]. The duration of studies ranged from two [37] [48] to 18 months [52] (Mdn = 6 months). Only six studies reported separate figures for intervention and follow-up durations, with intervention duration ranging from three [51] to 10 months [46] (Mdn = 6 months) and the duration of follow-ups ranging from six [46] to 12 months [52] (Mdn = 7 months). Mode of delivery (Q = 1.17, p = 0.556) and the duration of study (β = 0.14, p = 0.535), intervention (β = − 0.04, p = 0.916) and follow-up (β = − 0.25, p = 0.574) were not significantly related to intervention effectiveness (see Table 3).

Different people delivered the health coaching interventions. In four studies, the health coaching intervention was delivered by untrained personnel [34, 41, 44, 46, 53], while the remaining 16 interventions reported training of the interventionist on health coaching. Seven studies relied on nurses to deliver coaching sessions [34, 36, 47‐49, 52], four studies provided interventions by trained health coaches [35, 37, 50, 51], and only one study was delivered by health coaches certified by the International Coach Federation (ICF) [45]. The remaining interventions were delivered by different professionals, including dental care providers [46], community health workers [36], dieticians [53], medical staff [38, 42], pharmacists [44], psychologists [43], college students [41], peer patients [40], and physicians [48]. Type of intervention provider was not significantly related to intervention effectiveness (Q = 1.24, p = 0.538) (see Table 3).

The heterogeneity of interventions was evident in relation to the employed approaches and underpinning theories. Out of the 20 papers, five studies did not report the use of theories [34, 37, 44, 48, 51, 53]. The remaining 15 were grounded in different theories or frameworks. Most studies employed motivational interviewing [35, 36, 40, 42, 45‐47, 49, 52], two studies used the transtheoretical model [38, 49], and self-efficacy theory, cognitive-behavioural therapy and social-cognitive theory were each used once [39, 46]. The use of theory was not significantly related to intervention effectiveness (Q = 1.34, p = 0.247), nor was the specific use of MI (Q = 0.23, p = 0.632) (see Table 3).

A total of 23 BCTs were identified across the 20 studies reviewed (see Table 5). Interventions were varied in terms of the number of BCTs that were utilized in each intervention, ranging from 0 to 9 BCTs. The median of BCTs used across all interventions was 5. The most frequently coded BCT was 1.1 goal setting (behaviour), which has identified in 13 interventions [34‐36, 38‐41, 45, 46, 49‐51]. 1.2 problem solving was the second most commonly identified BCT, reported in 10 interventions [35‐39, 41, 43, 49, 52, 53]. Two BCTs, 1.4 action plan [34, 35, 39, 40, 45, 46, 50, 53] and 3.1 social support (unspecified) [35, 37‐39, 44, 45, 47, 48], were each reported in eight studies. 1.7 review outcome goals, 1.8 behavioural contract, 2.2 feedback on behaviour, 4.1 instruction on how to perform a behaviour, 8.7 graded tasks, 12.5 adding objects to the environment, and 2.5 monitoring outcome(s) of behaviour by others without feedback were each used once in six interventions [37, 39, 46, 48, 52, 53]. No BCTs were identified in one study [42].

An overview of the use of different BCTs and effect sizes found in each study is presented in Table 5. The most effective intervention based on the effect size (d = 0.78) used only one BCT: 3.1 social support (unspecified) [44]. Only one BCT, 1.1 goal setting (behaviour,) was used across all the interventions with a medium effect size, although it was also the most commonly used BCT across interventions with small or trivial effects.

There was no evidence of an association between the number of BCTs used in an intervention and its effect size (β = − 0.11, p = 0.651) (see Table 2). Of the moderation analysis with 23 different BCTs identified, only two analysis yielded significant results. Specifically, interventions that used credible sources of information (BCT 9.1) (Hedges’ g₊ = 0.08, k = 5) were significantly less effective than interventions that did not use this BCT (Hedges’ g₊ = 0.34, k = 15; Q = 7.67, p = 0.006). In addition, interventions that used social reward (BCT 10.4) (Hedges’ g₊ = 0.01, k = 3) were significantly less effective than interventions that did not use this BCT (Hedges’ g₊ = 0.32, k = 17, Q = 3.92; p = 0.048).

Although some studies showed good methodological quality due to their low bias [44, 45, 50‐52], the majority were weak because of either high or unclear risk of bias [34, 35, 37‐43, 46‐49, 53]. Eleven of the 20 studies [34, 39, 42, 44, 45, 47, 49‐53] described the method of randomization generation and 10 studies [34, 40, 42, 44, 45, 47, 50‐53] used a concealed allocation schedule. The methodological quality of blinding participants and personnel on the assignment of participants to study groups were generally low due to either high or unclear bias in procedures across most studies and insufficient detail. Across all the included studies, attrition bias and selective outcome reporting bias were low and not detected. Table 4 and Fig. 2 provide further details about the quality of the included studies.

This review has various strengths. First, it is the first review to identify the use of BCTs in health coaching studies with T2DM. Second, this review conducted a meta-analysis to investigate and evaluate the effectiveness of the BCTs in health coaching interventions and whether there is a link between using specific BCTs and reductions in Hb1Ac. Third, using the BCTs taxonomy assisted in systematically investigating and analysing interventions’ descriptions to identify the active ingredients of each intervention.

Additionally, there are several limitations to this review as well, which should be mentioned. First, it was limited to only English language papers, hence there is a possibility that some health coaching RCTs were not included. Second, studies have used various BCTs with different outcome measures, so it was difficult to determine which BCT assigned to HbA1c as an outcome. Consequently, it was difficult to be assured whether the positive results were achieved by individual BCTs or due to combinations of different BCTs. Inadequate reporting of intervention details and imprecise descriptions could lead to incorrect assumptions about the presence or absence of BCTs. Clarity and the amount of provided details on the interventions play a crucial role in coding BCTs correctly and so may have limited the accuracy of coding in the current review.