Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

A stepped wedge cluster randomised trial of nurse-delivered Teach-Back in a consumer telehealth service

  • Suzanne Morony ,

    Roles Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Project administration, Visualization, Writing – original draft, Writing – review & editing,

    Affiliation Sydney School of Public Health, The University of Sydney, Sydney, Australia

  • Kristie R. Weir,

    Roles Conceptualization, Investigation, Methodology, Resources, Writing – review & editing

    Affiliations Sydney School of Public Health, The University of Sydney, Sydney, Australia, Health Literacy Lab, Sydney School of Public Health, The University of Sydney, Sydney, Australia

  • Katy J. L. Bell,

    Roles Data curation, Formal analysis, Writing – review & editing

    Affiliation Sydney School of Public Health, The University of Sydney, Sydney, Australia

  • Janice Biggs,

    Roles Conceptualization, Resources, Writing – review & editing

    Affiliation Healthdirect Australia, Sydney, Australia

  • Gregory Duncan,

    Roles Methodology, Resources, Writing – review & editing

    Affiliation Eastern Health Clinical School, Monash University, Box Hill, Australia

  • Don Nutbeam,

    Roles Supervision, Writing – review & editing

    Affiliations Sydney School of Public Health, The University of Sydney, Sydney, Australia, Health Literacy Lab, Sydney School of Public Health, The University of Sydney, Sydney, Australia

  • Kirsten J. McCaffery

    Roles Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – review & editing

    Affiliations Sydney School of Public Health, The University of Sydney, Sydney, Australia, Health Literacy Lab, Sydney School of Public Health, The University of Sydney, Sydney, Australia



To evaluate the impact of Teach-Back on communication quality in a national telephone-based telehealth service, for callers varying in health literacy.


Cross-sectional stepped wedge cluster randomised trial with continuous recruitment, short (fixed) exposure and blinded outcome assessors. Nurses were stratified by hours worked and randomised into training groups using a computer generated sequence.


An Australian national pregnancy and parenting telephone helpline.


Complex intervention involving a single 2-hour group Teach-Back training session, combined with ongoing nurse self-reflection on their communication following each call and each shift.


Participants were 637 callers to the telephone helpline aged 18–75 (87% female), of whom 127 (13%) had inadequate health literacy (measured with the Single Item Literacy Screener); and 15 maternal and child health nurses with 15 years’ experience on average.


Primary outcome was a modified subscale of the Health Literacy Questionnaire, ‘having sufficient information to manage health’. Secondary caller outcomes included caller confidence, perceived actionability of information and nurse effort to listen and understand. Nurse outcomes were perceptions of their communication effectiveness.


Over a 7 week period, 376 surveyed callers received usual care and 261 Teach-Back. Ratings on the primary outcome increased over time (OR 1.17, CI 1.01 to 1.32, p = 0.03) but no independent Teach-Back effect was observed. A consistent pattern suggests that, compared with usual care, Teach-Back helps callers with inadequate health literacy feel listened to (OR 2.3, CI 0.98 to 5.42, p = 0.06), confident to act (OR 2.44, CI 1.00 to 5.98, p = 0.06), and know what steps to take (OR 2.68, CI 1.00 to 7.17, p = 0.06). Nurse perceptions of both their own communication effectiveness (OR = 2.31; CI 1.38 to 3.86, p<0.0001), and caller understanding (OR = 2.56; CI 1.52 to 4.30, p<0.001) both increased with Teach-Back. No harms were reported.


Teach-Back appears to benefit telephone health service users with inadequate health literacy, but the extent of this is unclear due to smaller numbers of lower literacy participants. Improving caller ratings over time are likely due to increasing nurse communication competence.



Telephone based telehealth services offer a convenient, cost-effective, and accessible way to provide health information and instructions[1] yet also carry the risk that callers may misunderstand the information they have been given[2]. (Even in face-to-face consultations) patients may misunderstand or forget what they have been told or fail to adhere to instructions [35]. This is particularly problematic for people with low health literacy, who are less likely to ask questions or indicate when they do not understand [612]. Telephone health providers can send supplementary written information electronically; however, for people who have difficulties reading (in English) or accessing information online this may be insufficient support. Electronically mediated health services including telehealth may have potential to worsen health disparities if people at risk of social health inequalities lack knowledge of telehealth services and how and when to use them[13].

A key communication technique to improve patient understanding and safety is Teach-Back [4, 5, 1417]. Teach-Back is a process whereby the health information provider (e.g. nurse) takes on the burden of ensuring effective communication e.g. by saying “I want to make sure I have explained everything correctly”, and asks the patient to restate the information or advice they have been given, using their own words [18]. The clinician can then correct any misunderstandings or add reminders of important actions to take if required [18]. Previous randomised controlled trials have found Teach-Back effective in improving emergency department discharge [19] and obtaining informed patient consent [20, 21]. Teach-Back has been used in outpatient management [16], to educate, assess learning and improve recollection of health information [22], and improve accuracy of laboratory results communicated between professionals by telephone [23]. It was recently ranked the number one health literacy practice by experts [17] and may also help health services improve client satisfaction and meet expectations [24].

Despite recommendations to use Teach-Back for telephone consultations, [1] [2] to our knowledge there are no previous efforts to systematically evaluate Teach-Back in a telephone based telehealth service. Teach-Back has sometimes been difficult to implement in health services, largely due to perceptions that it can increase consultation time, healthcare provider discomfort and concerns with phrasing, and quantity and complexity of information to discuss. We have discussed these issues and detailed the intervention we designed to address them elsewhere [25].

We aimed to implement Teach-Back in a telephone service providing information and advice for pregnancy and parenting of young children, and evaluate the impact on caller ratings of information they received and experience of the call. Trial and training session duration were constrained, and it was desirable for all participating nurses to receive the intervention. We designed a complex intervention to address key barriers to using Teach-Back and maximise nurse self-directed learning (see[25]) which we implemented in a stepped wedge RCT design. We have reported nurse and caller experiences with Teach-Back elsewhere[26].



The Pregnancy Birth and Baby (PBB) helpline is a government-funded national telephone service offering free information and advice on pregnancy and parenting (up to 5 years), staffed by qualified maternal and child health nurses at a single call centre. Nurses use computers to record caller data and access and forward online information; but can leave their desks using wireless headsets.

Study design

We conducted a cross-sectional stepped wedge cluster randomised trial with three-steps, continuous recruitment, and short (fixed) exposure [27]. Participants were callers to the helpline, recruited by helpline nurses (clusters). The intervention was nurse training in Teach-Back (detailed in [25]), and outcomes (impact of Teach-Back) were self-report measures of callers and nurses.

Initially all nurses were in the unexposed condition; following Teach-Back training, nurses crossed over to the exposed condition (Fig 1). This design was chosen for pragmatic reasons: firstly so that all participating nurses could be trained in Teach-Back; and secondly to minimise the crossover period and possibility of contamination.

Trial registration

Australian New Zealand Clinical Trials Registry, ACTRN12616000623493. Retrospectively registered.

Patient involvement

The trial was developed as part of a wider health literacy initiative at Healthdirect Australia, aiming to improve patient care. Survey questions were piloted with a small (highly educated) convenience sample of parents and expecting parents. Study participants were invited to receive study feedback by email.

Piloting and field testing

Several weeks prior to study commencement JB and DZ listened to a subset of calls and confirmed that Teach-Back was not currently being used, and that calls shorter than 4 minutes duration were unsuitable for Teach-Back. Following piloting (see above), the survey was field tested by a third party survey company (Population Research Laboratory (PRL)) with a sample (n = 20) of helpline callers.



Consenting callers were eligible if they were at least 16 years old and fluent in English. Exclusions were calls shorter than 4 mins duration, sensitive topic (e.g. pregnancy or child loss, family violence), caller distress, nurse not participating, or caller transferred to other services. Telephone numbers were screened by PRL to exclude callers already recruited into the study.


All nurses who were rostered to work on average more than 3 hours per week during the study period (n = 18) were invited to participate.


For callers we obtained informed oral consent at two points. This procedure was approved by the University of Sydney Human Research Ethics Committee (2016/083) and accepted by the Royal District Nursing Service Ethics Committee (approval number 150013). Oral consent to be contacted for follow-up was recorded using the same system regularly used by Healthdirect Australia to obtain consent for surveying client satisfaction. We amended the consent script to make it clear that the data collected during this study would be used for research purposes. Consent for follow-up is recorded on IT systems managed by Healthdirect Australia. Caller contact details for follow-up were only available to the researchers for those callers who had provided oral consent.

Approximately one week after their initial helpline call, consenting callers were contacted by trained female telephone interviewers and invited to participate in the research. Before the interview could commence callers were asked to consent to be interviewed. They were informed the survey was voluntary and their information would be kept confidential.

Written nurse consent forms were distributed and completed electronically.


Training in theory and skills for using Teach-Back was a 2-hour “communication skills” workshop (7–8 nurses per group), 2 or 4 weeks after study commencement. During the workshop trainers and nurses discussed how to use Teach-Back for different kinds of calls at the helpline, and acknowledged that not all calls would be suitable for Teach-Back. For example, some callers may simply want to confirm with a nurse that a course of action they have decided on (e.g. take a child to the doctor) is appropriate; Teach-Back would be unnecessary and inappropriate in such situations. For the duration of the study, nurses were encouraged to reflect after each call on how effectively they communicated and how well the caller understood. This is detailed in [25] and summarised in Fig 1.Emails encouraging participation and reiterating study procedures were sent to both nurse groups, in addition to individualised responses to any questions.



Caller outcomes were assessed in a single telephone survey conducted by PRL approximately one week following initial contact. The primary outcome was “having sufficient information to manage health”, a subscale of the Health Literacy Questionnaire (HLQ) [28]. This scale is comprised of 4 likert-type items with 4 response categories. We used a modified version that replaced references to “health” with terminology suited to pregnancy and parenting issues (S1 Text). Secondary outcomes are caller satisfaction [29], confidence to manage the issue (adapted from [30] [31]), actionability of advice (developed from the Patient Education Materials Assessment Tool [32] actionability component), expectations [29], self-reported adherence to advice, and feeling listened to and understood (2 items from the CollaboRATE measure of shared decision-making[33]). Several measures were adapted for telephone by capitalising words to be emphasised and moving scale anchors for consistency across scales. A complete list of variables and scales is detailed in S1 Table. To examine the impact of Teach-Back for participants with low health literacy, we assessed callers’ health literacy level using the Single Item Literacy Screener (SILS) [34], which asks how frequently adults need help reading written health material. Health literacy was dichotomised as adequate (never, rarely, sometimes)/inadequate (often, always).


Nurse measurements were two self-reported items on a five-point scale: “How effective do you think you were at communicating the information” and “How well do you think the caller understood your instructions or recommendations?”; with responses recorded in an online survey following each call (see [25]). In the Teach-Back condition, surveys also captured whether nurses reported using Teach-Back[25].

Sample size

Calculations indicated a sample of 400 callers (200 receiving Teach-Back and 200 receiving usual care), would be sufficient to detect a difference of at least 0.5 standard deviation (SD)[35] in the mean total score for the HLQ subscale ‘having sufficient information to manage health” [28, 36] with 80% power at the two-sided 5% significance level, assuming 16 clusters (nurses), a minimum of 5 participants (callers) per cluster per step, unequal cluster size, estimated 80% response rate and intention to treat analysis. To mitigate against possible nurse dropout or non-participation, we oversampled, aiming for a total sample of n = 600.

The trial was designed to run over 6 weeks. Numbers recruited in the usual care condition were higher than anticipated so the trial was extended by an additional week to balance the groups.

Randomisation and blinding


Callers to the PBB Helpline (participants) were invited by the nurse who took their call to participate in a survey for quality, service improvement, and research purposes. The Teach-Back trial was not explicitly disclosed to callers. Nurses used the same recruitment script in both conditions. Callers were allocated to condition according to whether the recruiting nurse had attended Teach-Back training (Fig 1). Caller interviewers (PRL) were blinded to the study design, allocation, and intervention.


Eligible nurses were stratified (into 4 strata) based on average number of hours rostered per week (to balance the number of calls to each condition), and randomly allocated to one of two training groups by SM using the RAND function in Excel. Nurses were invited to participate via a video introduction to the study and online consent form. They were advised they had been randomised into training groups for a “communication skills workshop” on a specific date. To minimise contamination nurses were asked not to discuss the content of training outside of their group and to avoid sharing training materials until both groups had completed training.

Statistical analysis

Analysis was by intention to treat. Due to highly skewed distributions of both caller and nurse data on most non-categorical outcomes we dichotomised all outcome variables into “highest scale category” and “other”.

We operationalised effects of time as week (1–7) of initial call to the helpline (measured as date of call for caller outcomes or date of self-reflection for nurse outcomes) and examined the data for an underlying time trend [37]. We used non-linear mixed effects models (logistic regression; NLMIXED procedure in SAS 9.4) to estimate the effects of Teach-Back on multiple outcomes that reflected (i) caller perceptions of advice (single measurements per caller) (ii) nurse perceptions of communication (repeated measurements per nurse over the course of the trial).

Models for the caller data were adjusted for time and health literacy, and included random intercepts for nurses to account for clustering of callers [37, 38]. We used Likelihood Ratio Tests and Z-scores to test for statistical significance of main effects and pre-specified subgroup analyses within health literacy subgroups respectively. Confidence intervals were calculated using the models’ beta coefficients and standard errors, assuming a normal distribution on the log-scale, with back-transformation to the natural scale (i.e. LCL = exp(beta coeff –1.96×SE) and UCL = exp(beta coeff+1.96×SE)). The intra cluster coefficient (ICC) was calculated as (variance of random intercepts / [variance of random intercepts + π2/3]). We analysed whether Teach-Back effects on caller perceptions differed between nurses (Teach-Back random effects) and if they changed over time (Teach-Back × time interactions), allowing for covariance parameters with the random intercepts (baseline differences between nurses), but set this to zero if required for model convergence.

For nurse self-reflections, we included random intercepts for nurses to account for clustering of repeated measurements within nurses and tested for both fixed and random effects of time and Teach-Back, and for Teach-Back × time interactions.


Participant flow

Eighteen eligible nurses were identified and randomised to training groups; three did not participate due to rostering or leave. Nurses recruited 1588 callers between 26 April and 15 June, 2016, of which 1215 were submitted to the survey company (Fig 2). After removing duplicates and wrong numbers there were 1025 eligible calls, for which 666 caller surveys were completed. Of these, 29 were excluded because the recruiting nurse did not participate in the study. The overall response rate was calculated as 65.5%, taking into account partially completed surveys (n = 5, not included in analyses). This is acceptable [39]. In the final sample (n = 637), 376 were in the usual care (59% completion rate) and 261 in the Teach-Back condition (49% completion rate). These completion rates are underestimates and do not account for the 190 calls that were duplicates or wrong numbers; we do not know allocation for those calls.


Recruitment of callers to the study was highly variable across nurses (detailed in Table 1) and reflected differences in the case load for each nurse. The number of eligible callers decreased each week as repeat callers were excluded. Two nurses recruited fewer than 15 surveyed callers, and more than half the participating nurses (n = 9) recruited fewer than 5 surveyed callers at each “step” in the study. The coefficient of variation in cluster size (CV)[40, 41] (standard deviation of cluster size divided by mean cluster size), was 0.89. Time between initial helpline call and follow-up survey ranged from 2–38 days (mean 9; median 8); IQR = 4.

Table 1. Number of cases (callers) recruited at each stage of the trial.

Shading indicates the clusters (nurses) were exposed to the intervention.

Demographics and descriptive statistics

Caller characteristics are detailed in Table 2. Callers were predominantly female (87%), university educated (64%) and had adequate health literacy (80%). Almost half (48%) identified as the mother of the subject of the call, 12% as a father or male partner, and 36% as the subject. Call topics could be broadly classified as antenatal (30%) and postnatal (70%) issues.

Table 2. Caller demographic characteristics for whole sample and inadequate (low) health literacy.

Around 33% of callers spoke a language other than English at home, which is higher than 2016 population estimates (22%)[42]. Of these, 81% had attended university, possibly reflecting greater service uptake among skilled migrants. Given the high associations between education and language spoken at home, we did not conduct planned subgroup analyses for language spoken at home.

The 15 participating nurses (clusters) were all female, aged 37–61 years (mean 54.9) and had been registered as nurses for 13–41 years (mean 28.8 years) and working in maternal and child health for between 7 months and 33 years (mean 15 years). Two had worked on another helpline previously. Workload per week per cluster ranged from a single shift (3–4 hours) to 4 days (32 hours). Demographic information is unavailable for 3 nurses who worked a small number of shifts.

Caller outcomes

Descriptive statistics for the dichotomised responses on each caller outcome in Teach-Back and usual care groups, both overall and by health literacy subgroup, are presented in Table 3. These indicate that ratings of the helpline were high overall; and that within conditions (i.e. usual care or Teach-Back) ratings tend to be higher among participants with adequate health literacy. For the primary outcome “having sufficient information to manage health” (dichotomous scale) the ICC was calculated at 0.019. The results of the mixed effects models analysis (Table 3) are summarised below: temporal trends first, followed by Teach-Back.

Table 3. Descriptive and model results for caller outcomes. All analyses adjusted for time and health literacy.

Temporal trend.

Caller perceptions of information sufficiency (primary outcome) increased each week as the trial progressed (OR 1.17, CI 1.01 to 1.32, p = 0.03), as did Actionability of advice “the nurse checked that I fully understood the information” (OR 1.16, CI 0.99 to 1.37, p = 0.05). Caller expectations “was the information you were given what you expected to hear” also changed over time, (OR 0.86, CI 0.75 to 0.98, p = 0.03); qualitative responses indicated later callers found it better than expected. We found no other statistically significant effects of time.

Primary outcome–Having sufficient information to know what to do.

After accounting for the time trend we found no incremental main effect of Teach-Back (OR 0.77, CI 0.44 to 1.37, p = 0.37) on the primary outcome, and no differences between the health literacy subgroups (p = 0.44; Table 3).

Secondary outcomes.

After accounting for time we found no main effects of Teach-Back. For several outcomes there was some evidence that callers with inadequate health literacy in the Teach-Back condition gave higher ratings compared to those receiving usual care; and modest evidence of a more positive impact of Teach-Back for callers with inadequate health literacy. This pattern of differences held for outcomes relating to: Confidence, “after speaking to the nurse, how confident were you in knowing how to manage the issue?” (OR = 2.44, 95% CI 1.00 to 5.98, p = 0.06 [Teach-Back vs usual care for participants with inadequate health literacy]; vs OR = 1.19, p = 0.10 [Teach-Back participants with adequate heath literacy]); Actionability, “the nurse told me all the steps I should take” (OR 2.68, 95% CI 1.00 to 7.17, p = 0.06; vs OR = 1.22, p = 0.10); and Shared decision-making, “how much effort was made to listen to the things that mattered the most to you regarding the problem that you called about” (OR = 2.30, 95% CI 0.98 to 5.42, p = 0.06; vs OR = 1.12, p = 0.09), and “how much effort was made to help you understand” (OR 1.97, 95% CI 0.83 to 4.66, p = 0.12; vs 0.86 for inadequate vs adequate health literacy, p = 0.048 for subgroup difference in effect). There was modest evidence suggesting that participants with inadequate health literacy in the Teach-Back condition were more likely to have called the helpline again (since the initial call) than those receiving usual care (OR 2.09, 95% CI 0.89 to 4.94, p = 0.10). We found no evidence of difference in outcomes between Teach-Back and usual care conditions for callers with adequate health literacy.

Ancillary analyses.

There was no evidence that the effect of Teach-Back differed between nurses nor that it changed over time for any of the outcomes (p>0.2 for random effects of Teach-Back, assuming no correlation with random intercepts; p>0.2 for all Teach-Back by time interactions). There was some evidence that the outcome “The nurse told me all the steps I needed to take” differed between nurses at different time points in the study (p = 0.03 for testing random effects of time, assuming no correlation with random intercepts) but this did not materially change the estimated effect of Teach-Back. There was no evidence that time effects differed between nurses for any of the other outcomes (all p > 0.2).

Nurse outcomes

The online self-reflection surveys (Table 4) gathered a total of 1861 responses over a period of 7 weeks; 1005 in usual care and 856 in Teach-Back. The two self-reflection questions were highly correlated (Pearson’s r = .88, p < .001). Completion of the survey was inconsistent, and number of call reflections per nurse ranged from 14 to 608. This is partly due to differences in the number of shifts each nurse worked during the trial period and differences in interpretation of the instructions to self-reflect. It may also reflect engagement with the study and perceived importance of self-reflection activities. The total number of responses per cluster per condition was generally balanced, although 3 clusters (<45 responses per cluster) exhibited a highly skewed response distribution (i.e. >80% of their self-reflections were in one condition).

Table 4. Effects of Teach-Back intervention on nurses (n = 15).

All analyses by intention to treat.

There was very strong evidence that Teach-Back improved nurse perceptions both that their communication was effective (OR = 2.31; CI 1.38 to 3.86, p<0.0001), and that the caller understood them (OR = 2.56; CI 1.52 to 4.30, p<0.001); the magnitude of these effects differed between nurses (p<0.001 for Teach-Back random effects for both outcomes, fitted allowing for correlation with random intercepts). There was also very strong evidence that the effects of Teach-Back increased with time (OR = 1.40 vs 4.69 and OR = 1.57 vs 6.56 for effects at weeks 3–4 vs 5–7 for effectiveness of communication and caller understanding respectively; p<0.001 for Teach-Back×Time interactions for both outcomes).

Contamination/carry over effects.

Some participants called the helpline more than once during the trial period, and could have received both Teach-Back and usual care. Although no caller was surveyed more than once, some repeat callers may have answered the survey questions with respect to a different call, meaning assignment to condition (Teach-Back or usual care) could be incorrect. We searched the database of eligible calls (n = 1215) for repeat phone numbers: almost 20% of callers (n = 220) called more than once during the study period, and usually spoke with a different nurse. Of 637 surveyed callers, 58 (9%) called more than once during the study period: 39 (6.1%) called twice, and 19 (3%) called 3 times or more (maximum 9). Time between (adjacent) calls ranged from 0 to 49 days (median 6.5).

In focus groups and interviews following the trial (reported in [25] [26]) nurses reported minimal contamination during the crossover period, also noting they rarely heard each other talking with callers. This suggests Teach-Back is unlikely to have been used in the usual care condition; however, evidence from nurse self-reports indicates Teach-Back was not always used in the Teach-Back condition. Its impact may therefore be underestimated in the data.


No harms to nurses or callers were identified in this study.


To our knowledge this trial is the first attempt to systematically evaluate Teach-Back in a telephone helpline, and it adds to the growing evidence base on the use of Teach-Back to improve healthcare communication, particularly for consumers with lower health literacy [19, 4345]. We used a structured intervention (nurse self-reflection and Teach-Back skills training—detailed in [25]) and randomised stepped wedge study design. The response rate was adequate (65.5%), although some assumptions of the power calculations were not met.

Descriptive statistics suggest a positive effect of Teach-Back on a number of caller outcomes, including the primary outcome; however, a strong temporal trend in the data accounted for much of this variance. Although there was no independent effect of Teach-Back on the primary outcome (having sufficient information to manage health) our analysis of secondary outcomes identified that Teach-Back can increase perceptions relevant to knowing what to do to manage health for callers with inadequate health literacy. This includes perceptions of good communication practice (i.e. feeling listened to and helped to understand), confidence to know what to do, and knowing what steps to take. There is potential to improve child health outcomes if Teach-Back can improve parent comprehension of information [46]. A small number of callers who participated in qualitative interviews expressed mostly positive experiences with Teach-Back (see [26] for detailed report). Nurses’ perceptions of their own communication effectiveness also increased with the use of Teach-Back.

Prior work has often reported that health providers can be reluctant to use Teach-Back due to perceptions it can increase consultation time [24, 43, 4749]. We did not report impact on call duration, because during the study period nurses were encouraged to focus on practising Teach-Back skills rather than meeting call duration targets, and skill learning may have temporarily inflated average call duration. Call duration is an important metric for call centres and it will be important to examine this in future work. Some nurses self-reported that Teach-Back helped to close an interaction with a caller[26]; previous work also indicated Teach-Back can help manage consultation time [50].

We estimate (from self-reflection data) that around 30% of calls in the Teach-Back condition did not involve Teach-Back, and in In focus groups and interviews (reported in [26] [25]) nurses commented that this was sometimes because they forgot, did not feel Teach-Back was appropriate, or did not know how to use Teach-Back for that kind of call. This is a reflection on the short time frame for training and limited opportunities for nurses to learn from each other during the study period. It also implies that the impact of Teach-Back is likely to be underestimated in these data.

Given that caller measurements were taken as nurses were learning new skills, the significant interaction between Teach-Back and time is likely due to a “rising tide effect” [51] rather than calendar time. Caller perceptions of the care they received was likely improved by nurses’ purposeful reflection on how well they explained things, how well they thought the caller understood, and on the communication techniques that worked well for them. Nurse practice in self-reflecting and using Teach-Back may account for improvements over time in both caller ratings (that information was sufficient, that the nurse confirmed their understanding, and information was better-than-expected) and nurse self-ratings.

Strengths and weaknesses of the study

This study was embedded into a quality and safety improvement activity, using a randomised design and concealed allocation with low potential for contamination. The stepped wedge design used is a strength in terms of sample size efficiency and practicality of implementation; however the trial did not meet all assumptions of the power calculations. Unlike previous work[19] we had no objective measure of client comprehension or recall of information (which varied on every call so it was not possible to assess it objectively and uniformly across the sample), and a time delay between the consultation and data collection, which may have contributed to a lack of statistical effect.

A weakness is the lack of a true measure of “usual care” at baseline and the short timeframe for data collection. All measurements were taken after nurses had commenced self-reflection, which appeared to have an independent effect on caller outcomes. Combined with practice or learning effects, it may have diluted the Teach-Back effect in the data. This effect would be further diluted by the fact that not every caller surveyed in the Teach-Back condition received Teach-Back. The lower survey completion rate in Teach-Back also reflects the shorter timeframe for follow-up. Future work that can address these shortcomings may produce a more true measure of the impact of Teach-Back.

Inadequate health literacy was operationalised with a measure of needing help reading health information that is moderately effective at identifying patients with limited literacy [52], and captures those callers who are unable to benefit from supplementary written information sent by email. The proportion of people identified as having inadequate health literacy (20%) is within range of previous Australian health literacy surveys using telephone recruitment (7–26% [53]; 24%, [54]); however the relatively small number of surveyed callers with inadequate health literacy made it more difficult to detect differences for this important subgroup.


Although the results of this study were not definitive, the balance of evidence reported here (and [25, 26]) supports the use of Teach-Back by health service providers in call centres of this nature. It may also have useful applications in other areas of telehealth. Consumers with lower health literacy in particular appear to benefit from Teach-Back and we have no evidence for any negative impact on caller outcomes, including satisfaction. The technique used to facilitate nurse development of Teach-Back skills (i.e. active self-reflection on their communication) appears to improve both provider and caller perceptions of the quality of communication. This might play a useful role in development of other telephone health communication skills and is a worthwhile topic for future studies. As discussed above, impact on call duration (and by extension impact on cost of service provision) could not be reliably ascertained and should be examined in future research.

It is an ongoing challenge for helpline providers to reach and support consumers with lower health literacy, who typically have poorer access to written health information [55, 56] and rely more on verbal reinforcement. They may also have poorer knowledge of available telehealth services and lack of confidence using the telephone [57]. In Australia, many regional and remote communities have limited access to mainstream health services due to distance, and stand to benefit from telehealth. Our study suggests that systematically implementing Teach-Back in a telephone based telehealth service may help to support people with low health literacy, and help reduce inequalities in the provision of telephone health information.

Unanswered questions and future research

Studies of this type (real-time, real world conditions) present methodological challenges in managing exposure and consistency in relation to the intervention, and in measuring change in primary outcomes. In a trial of longer duration, introducing a “transition” period at each step during which the cluster would be considered neither exposed not unexposed [51], and an objective measure of implementation fidelity [58] may help to isolate the specific effects of Teach-Back. This could also inform about the impact of Teach-Back on call duration.

We do not have objective data on whether callers understood correctly or followed through on recommended actions, and it is possible some callers may have acted incorrectly on information they misunderstood, or only acted on part of the information. Some previous work suggests callers may sometimes think they have acted in accordance with advice when in fact they did not [59, 60]. Future work to improve access to telehealth services for people with low health literacy could include teaching telephone communication skills to adult learners at risk of low health literacy (see e.g. [57]) to potentially improve uptake.


Teach-Back appears to have promise for improving communication of health-related information in a consumer helpline. More work is needed to explore whether the systematic use of enhanced communication techniques such as Teach-Back is effective in different telehealth contexts.

Supporting information

S1 Text. Adaptations to Heath Literacy Questionnaire (primary outcome).


S1 Table. Complete list of registered secondary outcome measures and scale reliability and inter-item correlations.



We would like to thank Dianne Zalitis (Healthdirect Australia), Maureen Wilkinson (RDNS), and Elizabeth Bruhn (RDNS) for logistic and operational support, and all the nurses and callers who participated. We are grateful to Dr Lamiae Azizi who did the sample size calculations and assisted with the trial design, Professor Les Irwig who consulted on the design, Dr Joanne Gale and Dr Patrick Kelly, who offered statistical support, and Dr Sunny Naicker who provided helpful feedback on an earlier version of this report. We would like to thank Christine Hanley and the telephone interviewers at The Central Queensland University Population Research Laboratory (PRL) for improving the survey flow through field testing, and for collecting the caller data.


  1. 1. Car J, Sheikh A. Telephone consultations. BMJ. 2003;326(7396):966–9. pmid:12727771
  2. 2. Leclerc BS, Dunnigan L, Cote H, Zunzunegui MV, Hagan L, Morin D. Callers' ability to understand advice received from a telephone health-line service: comparison of self-reported and registered data. Health Serv Res. 2003;38(2):697–710. Epub 2003/06/06. pmid:12785568; PubMed Central PMCID: PMCPMC1360910.
  3. 3. Richard C, Glaser E, Lussier M-T. Communication and patient participation influencing patient recall of treatment discussions. Health Expect. 2016:n/a-n/a. pmid:27868327
  4. 4. Sudore RL, Schillinger D. Interventions to Improve Care for Patients with Limited Health Literacy. Journal of clinical outcomes management: JCOM. 2009;16(1):20–9. PMC2799039. pmid:20046798
  5. 5. Volandes AE, Paasche-Orlow MK. Health literacy, health inequality and a just healthcare system. Am J Bioeth. 2007;7(11):5–10. Epub 2007/11/21. pmid:18027287.
  6. 6. Nouri SS, Rudd RE. Health literacy in the "oral exchange": an important element of patient-provider communication. Patient Educ Couns. 2015;98(5):565–71. Epub 2015/01/27. pmid:25620074.
  7. 7. Gaglio B, Glasgow RE, Bull SS. Do patient preferences for health information vary by health literacy or numeracy? A qualitative assessment. Journal of health communication. 2012;17 Suppl 3:109–21. Epub 2012/10/04. pmid:23030565.
  8. 8. Kountz DS. Strategies for improving low health literacy. Postgrad Med. 2009;121(5):171–7. Epub 2009/10/13. pmid:19820287.
  9. 9. Hawkins A. Health literacy: a potential barrier in caring for underserved populations. Dis Mon. 2010;56(12):734–‐40. pmid:21168579
  10. 10. Berkman ND, Sheridan SL, Donahue KE, Halpern DJ, Crotty K. Low health literacy and health outcomes: an updated systematic review. Ann Intern Med. 2011;155(2):97–107. Epub 2011/07/20. pmid:21768583.
  11. 11. Mårtensson L, Hensing G. Health literacy—a heterogeneous phenomenon: a literature review. Scandinavian Journal of Caring Sciences 2012;26:151–60. pmid:21627673
  12. 12. Easton P, Entwistle VA, Williams B. How the stigma of low literacy can impair patient-professional spoken interactions and affect health: insights from a qualitative investigation. BMC Health Serv Res. 2013;13:319–. PMC3751726. pmid:23958036
  13. 13. Latulippe K, Hamel C, Giroux D. Social Health Inequalities and eHealth: A Literature Review With Qualitative Synthesis of Theoretical and Empirical Studies. J Med Internet Res. 2017;19(4):e136. Epub 27.04.2017. pmid:28450271.
  14. 14. DeWalt D, Callahan L, Hawk V, Broucksou K, Hink A, Rudd R, et al. Health Literacy Universal Precautions Toolkit. In: (Prepared by North Carolina Network Consortium TCGSCfHSR, The University of North Carolina at Chapel Hill, under Contract No. HHSA290200710014.), editor. Rockville, MD.: Agency for Healthcare Research and Quality, 2010.
  15. 15. Australian Commission on Safety and Quality in Health Care. Consumers, the health system and health literacy: Taking action to improve safety and quality. Sydney: ACSQHC, 2013.
  16. 16. Schillinger D, Piette J, Grumbach K, Wang F, Wilson C, Daher C, et al. Closing the loop: physician communication with diabetic patients who have low health literacy. Arch Intern Med. 2003;163(1):83–90. Epub 2003/01/14. pmid:12523921.
  17. 17. Coleman C, Hudson S, Pederson B. Prioritized Health Literacy and Clear Communication Practices For Health Care Professionals. HLRP: Health Literacy Research and Practice. 2017;1(3):e91–e9.
  18. 18. Brega AG, Barnard J, Mabachi NM, Weiss BD, DeWalt DA, Brach C, et al. Health Literacy Universal Precautions Toolkit, 2nd Edition. 2015.
  19. 19. Griffey RT, Shin N, Jones S, Aginam N, Gross M, Kinsella Y, et al. The impact of Teachback on comprehension of discharge instructions and satisfaction among emergency patients with limited health literacy: A randomized, controlled study. Journal of Communication in Healthcare. 2015;Mar;8(1):10–21. pmid:26617669
  20. 20. Fink AS, Prochazka AV, Henderson WG, Bartenfeld D, Nyirenda C, Webb A, et al. Enhancement of surgical informed consent by addition of repeat back: a multicenter, randomized controlled clinical trial. Ann Surg. 2010;252(1):27–36. Epub 2010/06/22. pmid:20562609.
  21. 21. Wadey V, Frank C. The effectiveness of patient verbalization on informed consent. Can J Surg. 1997;40(2):124–8. Epub 1997/04/01. pmid:9126126; PubMed Central PMCID: PMCPMC3952974.
  22. 22. White M, Garbez R, Carroll M, Brinker E, Howie-Esquivel J. Is Teachback associated with knowledge retention and hospital readmission in hospitalized heart failure patients Journal of Cardiovascular Nursing 2013;Mar-Apr;28(2):137–46. pmid:22580624
  23. 23. Barenfanger J, Sautter RL, Lang DL, Collins SM, Hacek DM, Peterson LR. Improving patient safety by repeating (read-back) telephone reports of critical information. Am J Clin Pathol. 2004;121(6):801–3. Epub 2004/06/17. pmid:15198350.
  24. 24. Jager AJ, Wynia MK. Who gets a teach-back? Patient-reported incidence of experiencing a teach-back. Journal of health communication. 2012;17 Suppl 3:294–302. Epub 2012/10/04. pmid:23030577.
  25. 25. Morony S, Weir K, Duncan G, Biggs J, Nutbeam D, McCaffery K. Enhancing communication skills for telehealth: Development and implementation of a Teach-Back intervention for a national maternal and child health helpline in Australia. BMC Health Serv Res. 2018.
  26. 26. Morony S, Weir K, Duncan G, Biggs J, Nutbeam D, McCaffery K. Experiences with Teach-Back in a telephone health service. HLRP: Health Literacy: Research and Practice. 2017;1(4):e173–e81. Epub 10/10/2017.
  27. 27. Copas AJ, Lewis JJ, Thompson JA, Davey C, Baio G, Hargreaves JR. Designing a stepped wedge trial: three main designs, carry-over effects and randomisation approaches. Trials. 2015;16:352. Epub 2015/08/19. pmid:26279154; PubMed Central PMCID: PMCPMC4538756.
  28. 28. Osborne RH, Batterham RW, Elsworth GR, Hawkins M, Buchbinder R. The grounded psychometric development and initial validation of the Health Literacy Questionnaire (HLQ). BMC Public Health. 2013;13:658. Epub 2013/07/17. pmid:23855504; PubMed Central PMCID: PMCPMC3718659.
  29. 29. Spiby H, Green JM, Hucknall C, Foster HR, Andrews A. Labouring to better effect: studies of services for women in early labour The OPAL study (OPtions for Assessment in early Labour). In: Report for the National Co-ordinating Centre for NHS Service Delivery and Organisation R & D (NCCSDO), editor.: National Coordinating Centre; 2006.
  30. 30. Chew LD, Bradley KA, Boyko EJ. Brief questions to identify patients with inadequate health literacy. Fam Med. 2004;36(8):588–94. Epub 2004/09/03. pmid:15343421.
  31. 31. McCaffery KJ, Morony S, Muscat DM, Smith SK, Shepherd HL, Dhillon HM, et al. Evaluation of an Australian health literacy training program for socially disadvantaged adults attending basic education classes: study protocol for a cluster randomised controlled trial. BMC Public Health. 2016;16(454).
  32. 32. Shoemaker SJ, Wolf MS, Brach C. Development of the Patient Education Materials Assessment Tool (PEMAT): a new measure of understandability and actionability for print and audiovisual patient information. Patient Educ Couns. 2014;96(3):395–403. Epub 2014/06/29. pmid:24973195.
  33. 33. Elwyn G, Barr PJ, Grande SW, Thompson R, Walsh T, Ozanne EM. Developing CollaboRATE: a fast and frugal patient-reported measure of shared decision making in clinical encounters. Patient Educ Couns. 2013;93(1):102–7. Epub 2013/06/19. pmid:23768763.
  34. 34. Morris NS, MacLean CD, Chew LD, Littenberg B. The Single Item Literacy Screener: evaluation of a brief instrument to identify limited reading ability. BMC Fam Pract. 2006;7:21. Epub 2006/03/28. pmid:16563164; PubMed Central PMCID: PMCPMC1435902.
  35. 35. Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life: the remarkable universality of half a standard deviation. Med Care. 2003;41(5):582–92. Epub 2003/04/30. pmid:12719681.
  36. 36. Beauchamp A, Buchbinder R, Dodson S, Batterham RW, Elsworth GR, McPhee C, et al. Distribution of health literacy strengths and weaknesses across socio-demographic groups: a cross-sectional survey using the Health Literacy Questionnaire (HLQ). BMC Public Health. 2015;15:678. Epub 2015/07/22. pmid:26194350; PubMed Central PMCID: PMCPMC4508810.
  37. 37. Davey C, Hargreaves J, Thompson JA, Copas AJ, Beard E, Lewis JJ, et al. Analysis and reporting of stepped wedge randomised controlled trials: synthesis and critical appraisal of published studies, 2010 to 2014. Trials. 2015;16:358. Epub 2015/08/19. pmid:26278667; PubMed Central PMCID: PMCPMC4538923.
  38. 38. Hussey MA, Hughes JP. Design and analysis of stepped wedge cluster randomised trials. Contemp Clin Trials. 2007;28. pmid:16829207
  39. 39. Davern M, McAlpine D, Beebe TJ, Ziegenfuss J, Rockwood T, Call KT. Are Lower Response Rates Hazardous to Your Health Survey? An Analysis of Three State Telephone Health Surveys. Health Serv Res. 2010;45(5 Pt 1):1324–44. PMC2965507. pmid:20579127
  40. 40. Eldridge SM, Ashby D, Kerry S. Sample size for cluster randomized trials: effect of coefficient of variation of cluster size and analysis method. Int J Epidemiol. 2006;35(5):1292–300. pmid:16943232
  41. 41. Lauer SA, Kleinman KP, Reich NG. The Effect of Cluster Size Variability on Statistical Power in Cluster-Randomized Trials. PLoS One. 2015;10(4):e0119074. PMC4382318. pmid:25830416
  42. 42. Australian Bureau of Statistics. 2016 Census QuickStats. 2017.
  43. 43. Badaczewski A, Bauman LJ, Blank AE, Dreyer B, Abrams MA, Stein REK, et al. Relationship between Teach-back and Patient-Centered Communication in Primary Care Pediatric Encounters. Patient Educ Couns. 2017.
  44. 44. Yin HS, Shaikh U, Kimball E, Randolph G, Mistry K, Healy J, et al. Outcome-based Continuing Education: A Quality Improvement Intervention to Improve Family-centered Communication in Pediatrics. Pediatrics. 2018;141(1 MeetingAbstract):153–.
  45. 45. Samuels-Kalow M, Hardy E, Rhodes K, Mollen C. "Like a dialogue": Teach-back in the emergency department. Patient Educ Couns. 2016;99(4):549–54. Epub 2015/11/26. pmid:26597382; PubMed Central PMCID: PMCPMC4808368.
  46. 46. Hironaka LK, Paasche-Orlow MK. The implications of health literacy on patient–provider communication. Arch Dis Child. 2008;93(5):428–32. pmid:17916588
  47. 47. Centrella-Nigro AM, Alexander C. Using the Teach-Back Method in Patient Education to Improve Patient Satisfaction. J Contin Educ Nurs. 2017;48(1):47–52. Epub 2017/01/19. pmid:28099678.
  48. 48. Schlichting JA, Quinn MT, Heuer LJ, Schaefer CT, Drum ML, Chin MH. Provider perceptions of limited health literacy in community health centers. Patient Educ Couns. 2007;69(1–3):114–20. Epub 2007/09/25. pmid:17889494; PubMed Central PMCID: PMCPMC2246059.
  49. 49. Ting X, Yong B, Yin L, Mi T. Patient perception and the barriers to practicing patient-centered communication: A survey and in-depth interview of Chinese patients and physicians. Patient Educ Couns. 2016;99(3):364–9. Epub 2016/01/19. pmid:26776708.
  50. 50. Duncan G, Emmerton L, Hussainy S, McNamara K, Stewart K, Swinburne G, et al. HeLP: Health Literacy in Pharmacy project. In: The Research & Development Program, editor.: Australian Government Department of Health,; 2015.
  51. 51. Hemming K, Haines TP, Chilton PJ, Girling AJ, Lilford RJ. The stepped wedge cluster randomised trial: rationale, design, analysis, and reporting. BMJ. 2015;350. pmid:25662947
  52. 52. Powers BJ, Trinh JV, Bosworth HB. Can this patient read and understand written health information? JAMA. 2010;304(1):76–84. pmid:20606152
  53. 53. Barber MN, Staples M, Osborne RH, Clerehan R, Elder C, Buchbinder R. Up to a quarter of the Australian population may have suboptimal health literacy depending upon the measurement tool: results from a population-based survey. Health Promot Int. 2009;24(3):252–61. pmid:19531559
  54. 54. Adams RJ, Appleton SL, Hill CL, Dodd M, Findlay C, Wilson DH. Risks associated with low functional health literacy in an Australian population. Med J Aust. 2009;191(10):530–4. Epub 2009/11/17. pmid:19912083.
  55. 55. Easton P, Entwistle VA, Williams B. How the stigma of low literacy can impair patient-professional spoken interactions and affect health: insights from a qualitative investigation. BMC Health Serv Res. 2013;13:319. Epub 2013/08/21. pmid:23958036; PubMed Central PMCID: PMCPMC3751726.
  56. 56. Graham S, Brookey J. Do Patients Understand? The Permanente Journal. 2008;12(3):67–9. PMC3037129. pmid:21331214
  57. 57. Morony S, Lamph E, Muscat D, Nutbeam D, Shepherd H, Smith S, et al. Improving health literacy through adult basic education in Australia. Health Promot Int. 2017.
  58. 58. Ferm L, Rasmussen CDN, Jørgensen MB. Operationalizing a model to quantify implementation of a multi-component intervention in a stepped-wedge trial. Implementation Science. 2018;13(1):26. pmid:29422080
  59. 59. Dale J, Crouch R, Patel A, Williams S. Patients telephoning A&E for advice: a comparison of expectations and outcomes. J Accid Emerg Med. 1997;14(1):21–3. pmid:9023617
  60. 60. Munro J, Nicholl J, O’Cathain A, Knowles E, Morgan A. Evaluation of NHS Direct first wave sites: Final report of the phase 1 research. 2001.