- Open Access
- Open Peer Review
What helps and hinders midwives in engaging with pregnant women about stopping smoking? A cross-sectional survey of perceived implementation difficulties among midwives in the North East of England
Implementation Science volume 7, Article number: 36 (2012)
Around 5,000 miscarriages and 300 perinatal deaths per year result from maternal smoking in the United Kingdom. In the northeast of England, 22% of women smoke at delivery compared to 14% nationally. Midwives have designated responsibilities to help pregnant women stop smoking. We aimed to assess perceived implementation difficulties regarding midwives’ roles in smoking cessation in pregnancy.
A self-completed, anonymous survey was sent to all midwives in northeast England (n = 1,358) that explores the theoretical explanations for implementation difficulties of four behaviours recommended in the National Institute for Health and Clinical Excellence (NICE) guidance: (a) asking a pregnant woman about her smoking behaviour, (b) referring to the stop-smoking service, (c) giving advice about smoking behaviour, and (d) using a carbon monoxide monitor. Questions covering Michie et al.’s theoretical domain framework (TDF), describing 11 domains of hypothesised behavioural determinants (i.e., ‘knowledge’, ‘skills’, ‘social/professional role/identity’, ‘beliefs about capabilities’, ‘beliefs about consequences’, ‘motivation and goals’, ‘memory’, ‘attention and decision processes’, ‘environmental context and resources’, ‘social influences’, ‘emotion’, and ‘self-regulation/action planning’), were used to describe perceived implementation difficulties, predict self-reported implementation behaviours, and explore relationships with demographic and professional variables.
The overall response rate was 43% (n = 589). The number of questionnaires analysed was 364, following removal of the delivery-unit midwives, who are not directly involved in providing smoking-cessation services. Participants reported few implementation difficulties, high levels of motivation for all four behaviours and identified smoking-cessation work with their role. Midwives were less certain about the consequences of, and the environmental context and resources available for, engaging in this work relative to other TDF domains. All domains were highly correlated. A principal component analysis showed that a single factor (‘propensity to act’), derived from all domains, explained 66% of variance in theoretical domain measures. The ‘propensity to act’ was predictive of the self-reported behaviour ‘Refer all women who smoke……to NHS Stop Smoking Services’ and mediated the relationship between demographic variables, such as midwives’ main place of work, and behaviour.
Our findings advance understanding of what facilitates and inhibits midwives’ guideline implementation behaviours in relation to smoking cessation and will inform the development of current practice and new interventions. Using the TDF as a self-completion questionnaire is innovative, and this study supports previous research that the TDF is an appropriate tool to understand the behaviour of healthcare professionals.
Smoking during pregnancy is a cause of fetal mortality, low birth weight, and preterm delivery  and increases the risk of congenital anomalies . In the United Kingdom, smoking causes an estimated 5,000 miscarriages, 300 perinatal deaths, and 2,200 preterm deliveries per year . In the northeast of England, 22.2% of women were smoking at the time of delivery in 2009/10 compared to a national average of 14.2% .
Midwives have a significant role in influencing pregnant women who smoke. In England, recent guidance on smoking in pregnancy from the National Institute for Health and Clinical Excellence (NICE)  recommended behaviours specifically for midwives, including using a carbon monoxide breath test, referring all pregnant women who smoke to stop-smoking services, and providing information to women about the risks to the unborn child.
It is important to understand whether and why current guidelines are not fully implemented and what drives the behaviour of health professionals in order to develop interventions to optimise guideline implementation and effectiveness of clinical care. There is currently no evidence about the extent to which guidelines are implemented by midwives in relation to smoking cessation. However, there is evidence that implementation of NICE guidelines varies between organisations, professional groups, and guidance types . A systematic review of qualitative and quantitative studies relating to interventions with pregnant women who smoke identified a number of factors that impeded the implementation of smoking-cessation interventions, such as concern about the potential negative impact on the relationship between the pregnant woman and midwife, and limited knowledge about the use of guidelines or protocols in practice . This latter factor was derived from studies of service providers in Australia, South Africa, and England.
Scientific theories of behaviour and behaviour change might contribute to our understanding of health professionals’ guideline-implementation behaviours by providing hypotheses and possible explanations for behaviour congruent with current evidence that can then be generalised across contexts . For example, many social cognitive theories of behaviour hypothesise that behaviour is a function of an individual’s intention to perform it and that this intention, in turn, is influenced by beliefs about the behaviour, such as the ease, social acceptability, and consequence of performing the behaviour . A systematic review found consistent and strong relationships between intention and clinical behaviours of healthcare professionals . These findings support the applicability of theoretical models of individual behaviour to health professional practice.
One limiting factor in using psychological theory to predict and explain behaviours has been the large number of overlapping psychological, organisational, and motivational theories of behaviour, which make it difficult to select theoretical models or constructs for implementation research. Michie et al. reviewed 128 theoretical constructs derived from 33 relevant theories identified as applicable to implementation science. They used an expert consensus approach to group these constructs based on their commonalities into 11 domains . These domains are (1) knowledge; (2) skills; (3) social/professional role and identity; (4) beliefs about capabilities; (5) beliefs about consequences; (6) motivation and goals; (7) memory, attention, and decision processes; (8) environmental context and resources; (9) social influences; (10) emotion; and (11) self-regulation/action planning. While drawing on theories with different traditions, definitions of behaviour, and intended applications, these domains broadly cover the full range of current scientific explanations for human behaviour hypothesised in current theories, without aiming to maintain the explanatory and causal status the individual constructs have in their respective theory. The aggregated nature of the theoretical domain framework (TDF) provides a pragmatic framework for the exploration of behavioural predictors. It has been used to investigate perceived difficulties encountered by health and social care professionals when implementing national guidelines using a focus group approach  and clinicians’ behaviour in relation to blood transfusions through semi-structured one-on-one interviews with consultants . Both studies concluded that using the TDF was particularly useful for initially exploring possible explanations for suboptimal implementation behaviours and suggesting the most suitable theories for a further investigation of these behaviours. Moreover, a systematic review of studies predicting clinician’s behaviour using social cognitive theories  used the TDF, plus other psychological factors, to categorise the variables used in these studies to predict intention and behaviour of health professionals. A recent study used a survey approach to identify implementation difficulties in delivering tobacco prevention and cessation counselling through dentists and dental hygienists in Finland . This study demonstrated the feasibility of identifying implementation difficulties through a survey based on the TDF and identified issues around environmental context and resources as the main implementation difficulty amongst dental providers. However, the relatively small and heterogeneous sample in the study, limited reliability of the domain survey measures, and the lack of an assessment of actual individual implementation behaviours limits the conclusions that can be drawn from this study.
Our study had two aims. Firstly, we aimed to investigate the perceived implementation difficulties of midwives, working in different roles and locations, in providing smoking-cessation advice to pregnant women who smoke. In line with Michie et al.’s TDF , we investigated perceptions that were in favour of, and in conflict with, giving advice as specified in the NICE guidance. Secondly, we sought insight into any relationship between the self-reported behaviour of referring women to smoking-cessation services and demographic and professional variables. This work was developed with the purpose of presenting the survey findings to midwives to help them determine what actions would support them in working more effectively with pregnant women who smoke. Results of this initiative will be reported in a separate paper.
We conducted a self-reported, anonymous cross-sectional survey.
Setting and participants
Participants were midwives employed in any of the eight acute National Health Service (NHS) hospital trusts in the northeast region of England, a region with around 30,000 deliveries per year.
The questionnaire was designed to assess the 11 psychological domains identified by Michie et al.. Questions were based on the behaviours recommended in the NICE guidance . These are (a) asking a pregnant woman about her smoking behaviour, (b) referring a pregnant woman to the stop-smoking service, (c) giving advice to a pregnant woman about her smoking behaviour, and (d) using a carbon monoxide monitor to assess a woman’s smoking status. These four behaviours were used once in relation to each domain. We used at least three questions relating to each domain to ensure that each domain was accurately assessed. The questionnaire used the evidence statements from a NICE systematic review  to inform the topic areas for the questions. In addition, other important behaviours identified as barriers to effective smoking-cessation advice, such as advising women to cut down rather than quit , were included. There were 47 questions overall. Questions on midwife behaviours and theoretical domains were randomly ordered in the final version. A free-text section was included at the end of the questionnaire for respondents to add comments on how midwives could best manage pregnant women who smoke or on issues raised by the questionnaire. Responses to each question were assessed on a five-point Likert scale, ranging from ‘strongly agree’ to ‘strongly disagree’. The questionnaire is available in Additional file 1. Table 1 shows the alignment of the evidence statements from the NICE systematic review and the domains. A summary of the constructs relating to each domain is shown in Table 2. We also collected data on age, smoking status, main place of work (e.g., community or delivery suite), employing trust, whether or not respondents had trained as a specialist in smoking cessation, and length of time respondents had been in midwifery practice.
Following the completion of the survey, we undertook a backward validation exercise with five reviewers (health psychologists and applied health scientists) who had a range of knowledge about the TDF. The results of the backwards validation are presented in Additional file 2. In five of the domains, all the questions were correctly matched with the intended domain. Four domains scored 69% or higher. Only half of the questions in the beliefs about consequences domain and 55% in the social influences domain were correctly matched by reviewers. This may explain why in this study there is a lack of distinction between the domains, leading to the high intercorrelations.
The paper-based, self-reported survey was undertaken during January and February 2011. All midwives in the region (n = 1,358) were asked to take part via an invitation letter from the head of midwifery in their trust. A reminder letter with another copy of the questionnaire was sent to all midwives, as non-responders could not be identified. They were offered an extension to the closing date after 10 days. On both occasions, a free-post return envelope was included. The questionnaire included no identifying details, and responses were therefore anonymous.
Questionnaires were double entered into a Microsoft Excel (Microsoft Corporation, Redmond, WA, USA) spreadsheet for data management and then exported to SPSS 17.0 (SPSS, Inc., Chicago, IL, USA) for analysis. Responses were scored from 1 (strongly disagree) to 5 (strongly agree). Where more than one response was indicated, the variable was coded as missing (for demographic variables) or as least-extreme response (for all other variables). Items worded negatively, such as ‘My attempts to discuss smoking with pregnant women are usually perceived as nagging’, were reverse-coded and had their wording changed for the presentation of the results to the format, ‘My attempts to discuss smoking with pregnant women are not usually perceived as nagging’. Mean scores were calculated for each question. Higher mean scores indicated greater agreement with the statement. An overall mean score was calculated for each of the 11 domains.
Some of the free text responses from midwives indicated that they did not have the use of a carbon monoxide (CO) monitor for measuring smoking status. In addition, in a related audit (unpublished) carried out in parallel with the survey, two trusts stated that not all midwives had access to a CO monitor. Furthermore, the seven questions about CO monitors consistently had high rates of missing responses (15% to 22%). Response rates for all other questions ranged from 94% to 99%. The questions relating to CO monitors were, therefore, excluded from the analysis. Delivery-unit midwives (n = 155) were excluded from the analysis after discussion with midwifery managers, since it was anticipated that they would not be involved in engaging women in discussion about smoking cessation or referring them to stop-smoking services.
Cronbach’s alpha was computed for each domain to assess the internal consistency of the questions in each domain. A linear regression analysis was undertaken to explore which variables were independently associated with the self-reported behaviour ‘I always refer pregnant women who smoke to a stop-smoking service’. This question was selected as the outcome variable because NICE guidance advises midwives to ‘Refer all women who smoke, or have stopped smoking within the last 2 weeks, to NHS Stop Smoking Services’, and referrals are the only behaviour suggested in the NICE guidance that is regularly recorded within each NHS trust. To circumvent possible problems with multi-colinearity in the regression analysis, we conducted a principal component analysis over the highly intercorrelated domain measures to assess whether a more parsimonious structure could describe the TDF data. Moreover, only those demographic variables that had statistically significant correlations with the dependent variable were entered in the regression model.
We tested the hypothesis that the relationship between demographic variables (years practiced as a midwife, age, training in smoking cessation, and main place of work) and self-reported referral behaviour was mediated by the TDF. Mediation was tested using the INDIRECT macro for SPSS , which estimated the path coefficients in the mediator model and generated bootstrap confidence intervals (resamples = 5000, bias-corrected and accelerated) for the indirect effects of main place of work on referring to the smoking-cessation service through the proposed mediator variable ‘propensity to act’. Additionally, length of time practiced as a midwife and training as a specialist in smoking cessation were included as covariates in the mediation model.
Ethical approval was granted by a proportionate review subcommittee of Sunderland NHS Research Ethics Committee (10/H0904/75). Research approval was granted by the Research and Development committees in all eight NHS trusts in the region.
Characteristics of the respondents
The overall response rate was 43% (n = 589). The number of questionnaires analysed was 364, following removal of the delivery-unit midwives. Table 3 shows the characteristics of participating midwives. Of these respondents, 75% (274) had worked as a midwife for 10 or more years and most (60%, n = 218) worked in the community or an integrated team.
Table 4 shows descriptive variables and correlations for all domains. Cronbach’s alpha values for the domain measures ranged from 0.61 (beliefs about consequences) to 0.87 (action planning). Nine out of 11 domains had a satisfactory alpha of >0.7. The remaining two were <0.7 but were considered adequate (knowledge = 0.68 and beliefs about consequences = 0.61) given the heterogeneous nature of the TDF.
Mean scores for all domains were above the value of 3.00, indicating generally favourable views of participants towards providing smoking-cessation advice to pregnant women who smoke. The two domains with the highest mean scores were ‘professional role and identity’ (4.31, standard deviation [SD] = 0.56) and ‘motivation and goals’ (4.28, SD = 0.57). The two domains with the lowest mean scores were ‘beliefs about consequences’ (3.25, SD = 0.63) and ‘environmental context and resources’ (3.48, SD = 0.79). All domains were highly correlated with each other and significantly but weakly correlated with length of practice as a midwife, age, and training in smoking cessation. Main place of work was moderately and significantly correlated with all domains. There was no correlation with smoking status.
Principal component analysis of the 11 domains identified one component with an eigenvalue of 6.601, accounting for 66% of the variability in TDF scores. This suggests that, in the present study, the psychological domains might best be described as a single factor. For all further analyses, we therefore constructed a grand mean for all 11 domains, which we called ‘propensity to act’, that describes the propensity of midwives to act in relation to pregnant women who smoke.
Table 5 shows that professional and demographic variables accounted for 10.6% of the variability in referral behaviour. Adding ‘propensity to act’ to this equation in a second step nearly doubled the predictive utility of the model to an adjusted R 2 = .199 (p < .001). ‘Propensity to act’ was the variable that was most strongly independently associated with behaviour, attenuating the beta value for main place of work considerably.
The hypothesised mediation model in Figure 1 shows that a significant indirect relationship between main place of work and the behaviour ‘refer pregnant women to smoking-cessation service’ acts via ‘propensity to act’ (beta [B] = −.42, standard error [SE] = .08 [95% confidence interval: –.60, –.27]). Of the behavioural variance, 20% was accounted for by the hypothesised model. The arrows in Figure 1 reflect the model specification; they do not suggest causality. There was a significant association of main place of work with behaviour (B = −.84, SE = .14, p < .001; indicates higher probability of referring for midwives working in integrated teams or in the community), if the mediator was not included in the model. The strength of this association decreased when the mediator was included in the model (B = −.42, SE = .15, p = .05) but remained significant, indicating that there is a significant indirect effect via ‘propensity to act’ and also a direct effect of main place of work on behaviour (partial mediation). Both, the direct effect of main place of work on ‘propensity to act’ (B = −.50, SE = .05, p < .001) and the effect of ‘propensity to act’ on referring to the smoking-cessation service (B = .85, SE = .13, p < .001) were significant. Finally, there was no significant effect of control variables on behaviour (trained in smoking-cessation advice: B = .08, SE = .17, p = .64; length of time practiced as a midwife: B = −.10, SE = .15, p = .51).
Summary of main findings
In this survey of a sample of midwives from northeast England using the TDF, respondents mostly displayed favourable views to providing smoking-cessation advice to pregnant women who smoke, had high levels of motivation, and saw this as an integral part of their role. Midwives were less positive about the consequences of their actions in relation to smoking cessation and the environmental context and resources available to them. Using a single-factor solution from a principal component analysis of data items relating to 11 components of the TDF, midwives’ ‘propensity to act’ accounted for the largest proportion of variance and was independently associated with referring pregnant women who smoke to a stop-smoking service. Mediation analysis indicated that main place of work was directly related to referral behaviour, though ‘propensity to act’ had a mediating effect on the relationship.
We were surprised to find that, in this study, all 11 domains were highly correlated and best described as a single homogeneous measure. Amemori et al. found substantial relationships between domain measures in their survey of implementation difficulties in tobacco use prevention and cessation counselling with dental providers in Finland, and factor analysis suggested describing the measures along three dimensions (motivation, capability, and opportunity) . These findings reflect that the TDF domains are not a theory. They have been developed to try and encompass a broad range of different theories to arrive at a comprehensive framework that incorporates their main theoretical explanations for behaviour. Due to their descriptive and integrative nature, the domains are not independent constructs, which could explain why in this study the domains were highly correlated with each other. For the present example, ‘propensity to act’ is highly related to place of work. One can argue that midwives in integrated teams and community midwives experience more opportunities to intervene than those in fetal medicine units, day assessment units, antenatal clinics, or inpatient wards. They thus perceive more support, knowledge, and motivation and therefore develop better capabilities over time. This study provides a starting point to understanding implementation difficulties. In order to investigate the specific causal mechanisms of midwives’ implementation of NICE guidance for smoking cessation in pregnancy, further research is needed.
Strengths and limitations of the methods
Michie et al.’s TDF of behaviour change  has been used previously in face-to-face interviews to study implementation of a mental health guideline  and to understand clinicians’ blood transfusion behaviour . To our knowledge, this study is the first to measure health professionals’ perceived implementation difficulties along the 11 theory domains identified by Michie et al. by both self-completion questionnaire and a self-report of implementation behaviour, namely referral of smoking pregnant women to NHS Stop Smoking Services. The questionnaire approach has the advantage of greater reach and thus generalisability, enabling us to assess professional behaviour in a large population over a geographically large administrative area. The questionnaire was based on previous research that has also been used in assessing the implementation of NICE guidance. The behaviours assessed in our questionnaire were drawn from a systematic review that formed the basis of NICE guidance on smoking in pregnancy.
At 43%, the response rate could be considered low, and other cross-sectional surveys of midwives have response rates ranging from 38%  to 81% . These surveys were not directly comparable to this one since they were either in different settings (country or area of work) or the questions were focused on other issues. Due to the anonymous nature of the survey, we were unable to quantify the effects of non-response bias, for example, according to midwifery role or age. The survey distribution method in each NHS trust was determined by local midwifery managers and, therefore, subject to some variation. It is possible that not all midwives received copies of the questionnaire. Time limitations associated with funding meant that only one reminder was sent out, which may have limited response further. As usual in survey research, not all midwives completed all the questions, leading to some item response bias . In addition, the questionnaire was perceived as long and repetitive by some respondents. Relatively high mean scores were reported for all domains, indicating that no one disagreed or strongly disagreed with the statements. This could be related to social desirability bias, since it is likely that midwives were aware of national guidance and of their expected professional role in relation to smoking cessation. Two of the 11 domains did not have satisfactory alpha scores, though overall our scores were higher than those reported by Amemori et al. In their study, they reported Cronbach alpha scores for nine domains, five of which were <0.6.
Smoking cessation in pregnant women is affected by many interacting factors. These include the socioeconomic status of the pregnant woman, her personal context, and the nature of the relationship that develops between each pregnant woman and the midwife she sees during her pregnancy. This study has focused on only one discrete aspect of the many factors that determine whether or not a pregnant woman who smokes stops. This was a cross-sectional study; therefore, we could only establish association, not causality. The outcome measure was self-reported and, thus, prone to recall or desirability bias. However, it was deemed important at this stage of research to conduct an anonymous survey, as some of the questions on implementation difficulties (e.g., questions about knowledge) might otherwise be subject to social desirability. We therefore decided to rely on self-reports, as data linkage would have undermined the confidence of participants in the anonymous nature of this survey.
Relationship to existing knowledge
The NICE systematic review  reported that some midwives were concerned that they may be perceived as nagging if they gave advice about stopping smoking when pregnant . Our results were consistent with this, as only 19% of respondents agreed that discussing smoking with pregnant women was not usually perceived as nagging.
Whilst Abrahamsson et al.  reported that Swedish midwives suggested discussing smoking with a pregnant woman was a potential threat to their relationship, our respondents did not seem to identify this as a concern, since 79% disagreed with the statement, ‘Suggesting a woman stops smoking when she is pregnant will make our relationship awkward in the future’. Similarly, NICE evidence statement 7  identifies skills and knowledge as being limiting factors to implementing smoking-cessation interventions, yet in our sample, knowledge was the domain with the second highest score. This could be due to social desirability influencing the responses of our sample, or indeed, could indicate that participants perceived that their training needs were met.
NICE evidence statement 8  highlights time as a barrier to implementing smoking-cessation interventions. In our sample, 39% of respondents disagreed or strongly disagreed that they had enough time to ask women about their smoking behaviour, whilst 41% agreed or strongly agreed they had enough time.
Compared to the study of dental professionals in Finland, this survey had higher internal consistency (α = 0.50 to α = 0.64 for Amemori et al. , α = 0.61 to α = 0.87 in this study). In both studies, the domain ‘environmental context and resources’ had one of the lowest scores and ‘motivation and goals’ was in the top three highest scores for both studies. However, the results for other domains were different in the two studies. This could reflect the different populations that were studied, in terms of professional expertise, for example, or reflect the ability of the TDF to discriminate between the different perceived implementation difficulties among different healthcare professional groups.
Implications for clinicians or policy makers
Midwives are a key group of health professionals who can influence pregnant women. It is, therefore, important to understand what helps and hinders their behaviours and what can facilitate them being more effective when they intervene with pregnant women who smoke. Using the TDF has furthered our understanding of what facilitates and inhibits midwives’ behaviour in relation to smoking cessation and offered important pointers for the development of current practice and new interventions. For example, the lower scores for ‘beliefs about consequences’ suggest that midwives may require additional information about the effectiveness of brief interventions in relation to smoking cessation. There were also lower mean scores for the ‘environmental context and resources’ and ‘memory, attention, and decision making’ domains. This suggests that midwives may not have the CO monitors or information leaflets that they require when they are engaging with pregnant women who smoke. The results may also indicate a need for prompts in the process that help remind midwives to ask women about their smoking behaviour. It is also helpful to know that the high scores on the ‘knowledge’ and ‘professional role’ domains indicate midwives are confident that they have a sufficient understanding of what they need to know about smoking in pregnancy and that they consider addressing this topic is part of their work as midwives.
One aspect of the study has been to share the results with midwives in northeast England at a stakeholder workshop. Their interpretation of the results is currently being used to inform actions that they can take to support the work they do with pregnant women who smoke. This work is ongoing and will be reported at a future date.
This study supports previous research that has found that the use of the TDF offers an appropriate way to understand the behaviour of healthcare professionals. The challenge is to refine the questionnaire so that it can be used either to differentiate more clearly between domains or determine which are the key domains that influence the ‘propensity to act’. Then a shorter (and more acceptable to participants) tool can be developed that could be used more widely as the baseline for assessing the drivers for healthcare professionals’ behaviour, before designing interventions that will change behaviour.
Our findings advance understanding of what facilitates and inhibits midwives’ actions in relation to smoking cessation and will help with the development of current practice and new interventions. Using the TDF as a self-completion questionnaire is innovative, and this study supports previous research that the TDF is an appropriate way to understand the behaviour of healthcare professionals.
Royal College of Physicians: Passive smoking and children. 2010, A report by the Tobacco Advisory Group
Hackshaw A, Rodech C, Boniface S: Maternal smoking in pregnancy and birth defects: a systematic review based on 173 687 malformed cases and 11.7 million controls. Hum Reprod Update Advanced access published July 11. 2011,http://humupd.oxfordjournals.org/content/early/2011/07/09/humupd.dmr022.full,
Department of Health: Statistical release: Smoking at delivery - Quarter 3. 2010,http://www.dh.gov.uk/en/Publicationsandstatistics/Statistics/StatisticalWorkAreas/Statisticalpublichealth/DH_124342, /11, Table 4: Trend in smoking at time of delivery, by PCT and SHA,
National Institute for Health and Clinical Excellence: Public health guidance 26, How to stop smoking in pregnancy and following childbirth. 2010
Sheldon TA, Cullum N, Dawson D, Lankshear A, Lowson K, Watt I, West P, Wright D, Wright J: What’s the evidence that NICE guidance has been implemented? Results from a national evaluation using time series analysis, audit of patients’ notes, and interviews. BMJ. 2004, 329: 999-1004. 10.1136/bmj.329.7473.999.
Baxter S, Blank L, Guillaume L, Messina J, Everson-Hock E, Burrows J: Systematic review of how to stop smoking in pregnancy and following childbirth. Review 2: Factors aiding delivery of effective interventions. 2009, The University of Sheffield School of Health and Related Research
Hrisos S, Eccles MP, Johnston M, Francis JF, Kaner EFS, Steen IN, Grimshaw J: An intervention modelling experiment to change GPs’ intentions to implement evidence-based practice: Using theory-based interventions to promote GP management of upper respiratory tract infection without prescribing antibiotics. BMC Health Serv Res. 2008, 8: 10-10.1186/1472-6963-8-10.
Webb TL, Michie S, Sniehotta FF: Using theories of behavior change to inform interventions for addictive behaviors. Addiction. 2010, 105: 1879-1892. 10.1111/j.1360-0443.2010.03028.x.
Eccles MP, Hrisos S, Francis J, Kaner EF, Dickinson HO, Beyer F, Johnston M: Do self-reported intentions predict clinicians’ behaviour: a systematic review. Implement Sci. 2006, 1: 28-10.1186/1748-5908-1-28.
Michie S, Johnston M, Abraham C, Lawton R, Parker D, Walker A: Making psychological theory useful for implementing evidence based practice: a consensus approach. Qual Saf Health Care. 2005, 14: 26-33. 10.1136/qshc.2004.011155.
Michie S, Pilling S, Garety P, Whitty P, Eccles MP, Johnston M: Simmons: Difficulties implementing a mental health guideline: an exploratory investigation using psychological theory. Implement Sci. 2007, 2: 8-10.1186/1748-5908-2-8.
Francis JJ, Stockton C, Eccles MP, Johnston M, Cuthbertson BH, Grimshaw JM, Hyde C, Tinmouth A, Stanworth SJ: Evidence-based selction of theories for designing behaviour change interventions: Using methods based on theoretical construct domains to understand clinicians’ blood transfusion behaviour. Br J Health Psychol. 2008, 14: 625-646.
Godin G, Belanger-Gravel A, Eccles M, Grimshaw J: Healthcare professionals’ intentions and behaviours: A systematic review of studies based on social cognitive theories. Implement Sci. 2008, 3: 36-10.1186/1748-5908-3-36.
Amemori M, Michie S, Korhonen T, Murtomaa H, Kinnunen TH: Assessing implementation difficulties in tobacco use prevention and cessation counselling among dental providers. Implement Sci. 2011, 6: 50-10.1186/1748-5908-6-50.
Preacher KJ, Hayes AF: Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Methods. 2008, 40: 879-891. 10.3758/BRM.40.3.879.
Jones D: Exploring the value of smoking cessation advice. Br J Midwifery. 2003, 11: 4-
Clasper P, White M: Smoking cessation interventions in pregnancy: practice and views of midwives, GPs and obstetricians. Health Educ J. 1995, 54: 150-162.
Bowling A: Research methods in health. 2002, Open University Press, United Kingdom, 2
Arborelius E, Nyberg K: How should midwives discuss smoking behaviour in pregnancy with women of low educational attainment?. Midwifery. 1997, 13: 210-215. 10.1016/S0266-6138(97)80008-3.
Abrahamsson A, Springett J, Karlsson L, Hakansson A, Ottosson T: Some lessons from Swedish midwives’ experiences of approaching women smokers in antenatal care. Midwifery. 2005, 21: 335-345. 10.1016/j.midw.2005.02.001.
We thank all the midwives who completed the questionnaire; the heads of midwifery, Mr. C. Curry, Mrs. S. Ford, Mrs. A. Holt, Miss J. Mackie, Miss J. McNichol, Mrs. Y. Regan, Mrs. G. Wiggham, and Mrs. C. Wilkinson; Dr. S. Sturgiss and the members of the North East Clinical Innovations Team for Maternity and Newborn; Mr. P. Gellert, who computed the mediation analysis; members of the workshop advisory group, Mrs. J. Anderson, Dr. S. Mackenzie, Miss J. Mackie, Miss D. Orange, and Mr. M. Willmore; and Ms. G. Crimmens and Mrs. M. Harrison, who provided administrative support for the questionnaire and workshop. We also thank the three reviewers who provided constructive and helpful feedback on the original manuscript.
NHS North East funded the questionnaire and workshop. JB is funded by the Northern Deanery. The research was undertaken under the auspices of Fuse, the UK Clinical Research Collaboration Centre for Translational Research in Public Health, which is funded by the British Heart Foundation, Cancer Research UK, the Economic and Social Research Council, the Medical Research Council, and the National Institute for Health Research. FFS is wholly and MW is in part funded from this source. The views expressed here are those of the authors and not necessarily those of the funders.
The authors declare that they have no competing interests.
EM had the idea for the programme of work and, with JB, MW, and RB, developed the initial proposal. JB, VAS, and FFS designed the questionnaire, with input from MW and RB. JB co-ordinated data collection and management. JB, VAS and FFS analysed the data, and all co-authors contributed to interpretation. JB produced the first draft of the manuscript, and all co-authors commented on subsequent drafts and approved the final version for publication.
Authors’ original submitted files for images
Below are the links to the authors’ original submitted files for images.
About this article
Cite this article
Beenstock, J., Sniehotta, F.F., White, M. et al. What helps and hinders midwives in engaging with pregnant women about stopping smoking? A cross-sectional survey of perceived implementation difficulties among midwives in the North East of England. Implementation Sci 7, 36 (2012). https://doi.org/10.1186/1748-5908-7-36
- Pregnant Woman
- Smoking Cessation
- National Health Service
- Smoking Behaviour
- Theoretical Domain Framework