ARTICLE IN PRESS

Midwifery (2005) 21, 204211

www.elsevier.com/locate/midw

Guidelines on anaemia: effect on primary-care midwives in The Netherlands

Pien Offerhaus, MSc (Health Scientist)a,b, Margot Fleuren, PhD (Senior Researcher)c, Michel Wensing, PhD (Senior Lecturer)a,
a

Centre for Quality of Care Research (WOK), University Medical Centre, St Radboud Nijmegen, P.O. Box 9101, 6500 HB, Nijmegen, The Netherlands b Royal Dutch Organisation of Midwives (KNOV), Bilthoven, The Netherlands c TNO Quality of Life, Leiden, The Netherlands
Received 21 April 2004; received in revised form 8 September 2004; 23 September 2004; accepted 22 October 2004

KEYWORDS
Practice guidelines; Anaemia; Midwifery

Summary Objective: to assess the adherence and perceived barriers for implementation of a clinical-practice guideline on anaemia, which was the rst national guideline for primary-care midwifery in The Netherlands. Design: cross-sectional survey study. Setting: primary-care midwifery in The Netherlands. Participants: 160 midwives (60% response rate). Measurements: questionnaire on the knowledge of, and attitudes and self-reported adherence to, 14 key recommendations in the guideline; attitudes to guidelines in general; and perceived barriers to implementation. Findings: the number of midwives agreeing with and adhering to specic recommendations varied between 29 and 90%. Most midwives had a positive attitude to the guidelines. The most relevant general barriers were related to the behaviour of general practitioners and obstetricians (32% of the midwives reported this). Larger numbers of midwives mentioned barriers to specic aspects of the guideline, particularly alternative iron supplementation or dietary supplements (59%), and not prescribing iron supplementation if haemoglobin was low but mean corpuscular volume was normal (49%). Key conclusions: the guideline on anaemia was well received by primary-care midwives in The Netherlands, but implementation of specic recommendations needs further attention.

Corresponding author.

E-mail address: M.Wensing@wok.umcn.nl (M. Wensing). 0266-6138/$ - see front matter & 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.midw.2004.10.005

ARTICLE IN PRESS
Guidelines on anaemia: effect on primary-care midwives in The Netherlands 205

Implications for practice: the study provides evidence for the national organisation of midwives to continue with the development and implementation of clinical guidelines. & 2005 Elsevier Ltd. All rights reserved.

Introduction
Clinical-practice guidelines are systematically developed statements to assist practitioner and patient decisions about appropriate health care for specic clinical circumstances (Field and Lohr, 1990). They intend to translate research evidence and professional expertise on clinical management into specic recommendations for daily practice. In many countries, health professions and national organisations have developed clinical guidelines (Burgers et al., 2003). In November 2000, the Royal Dutch Organisation of Midwives (KNOV) launched its rst national guideline, which focused on the diagnosis and treatment of iron-deciency anaemia in pregnancy and after labour by primary-care midwives (AmelinkVerburg et al., 2000). The incidence of iron-deciency anaemia is 4% a year in non-pregnant women aged 2050 years (Looker et al., 1997). In The Netherlands, primary-care midwives diagnose and treat anaemia without the involvement of doctors. Anaemia was chosen as the topic for the rst guideline, because it was suspected, on the basis of observational studies (Wildschut et al., 1998), that haemoglobin (Hb) tests were carried out too often, and that iron supplementation was often used without indication. Improving test ordering and prescribing patterns was therefore expected to reduce unnecessary diagnostic procedures and prescription of medication, and thus reduce costs and increase efciency of health services. The intended impact of the guideline can only be expected if the professional behaviour of primary care-midwives is consistent with the recommendations made in the guideline. Insight into the knowledge, attitudes and (obstacles in) adherence of midwives to the guideline was lacking. In The Netherlands, the general opinion is that pregnancy and childbirth are natural physiological and not medical events. Maternity care is predominantly given in primary health-care settings. Antenatal care is mainly given by primary-care midwives who work in independent practices. A few general practitioners (GPs), mainly in rural areas, are also involved in maternity care. If, during pregnancy, the midwife suspects an increased risk of complications, the woman is referred to the obstetrics department in a hospital. If the preg-

nancy develops normally, the woman can choose to give birth at home or in the hospital, under supervision of her own midwife. After birth, the midwife continues her care for the mother and baby for the rst 710 days. Anaemia is one of the conditions that is routinely screened for during antenatal care. Primary-care midwives prescribe iron supplementation, which is an exception to the general principle that only doctors can prescribe a pharmaceutical treatment. A structured, evidence-based method was used to develop the guideline on anaemia. The guideline consisted of a review of the research evidence, a short guideline and a summary card (AmelinkVerburg et al., 2000). On the basis of their expertise and epidemiological knowledge, a panel of eight midwives was involved in developing the guideline in order to interpret the research evidence. An external panel of other professionals (among others, GPs and obstetricians) was asked for comments. In this guideline, new parameters to diagnose anaemia were dened. New cut-off points for Hb are recommended, adjusted for the variable level of physiological haemodilution during pregnancy. The lowest cut-off point is 6.3 mmol/l at 2237 weeks. The guideline advises the use of mean corpuscular volume (MCV) as an additional test to differentiate a physiological haemodilution from anaemia, because it is sensitive, cheap and easily available. The short guideline and a summary card were distributed to all midwives in The Netherlands in January, 2001. The national journal for midwives (Tijdschrift voor Verloskundigen) published a special issue on the guideline; all members of the national organisation of midwives (KNOV) receive this journal (more than 90% of all Dutch midwives). Other dissemination and implementation activities were planned, but not yet carried out at the time of this study. Dissemination of guidelines does not automatically result in optimal implementation in daily practice, because many factors may complicate the uptake (Grol and Grimshaw, 2003). In other words, reading the guideline does not necessarily change the professional routines of midwives. These include professional factors (e.g. negative attitudes, lack of specic skills, self-efcacy); patient/ client factors (e.g. inappropriate patient/client expectations, patient/client willingness to

ARTICLE IN PRESS
206 co-operate); characteristics of the guideline (e.g. complexity, relative advantage); and organisational and structural factors (e.g. inadequate organisational culture, staff turnover, existing rules and regulations). It is important to have insight into current practice and into impeding and facilitating factors for improvement, in order to design a targeted and effective implementation strategy that is adapted to these factors (Fleuren et al., 2004; Grol and Wensing, 2004). The aim of this study was to assess the adherence and perceived barriers to implementation of a clinical practice guideline on anaemia for primary-care midwifery in The Netherlands. P. Offerhaus et al. The questionnaire measured the following issues related to the 14 key recommendations. First, for each recommendation, the following areas were explored: self-reported knowledge of the recommendation (yes/no), opinion on the recommendation (agree, partly agree, disagree) and selfreported adherence to the recommendation (mostly, sometimes, rarely). Second, a validated instrument with 12 items on attitudes about guidelines in general, and 14 items on general barriers for implementation, was used (ve-point answering scale, agree-disagree format) (Peters et al., 2003). This instrument was based on a consensus procedure among implementation experts, reviews of studies on barriers for implementation and psychometric validation studies in primary-care doctors. In addition, 14 items (vepoint answering scale, agreedisagree format) were included on potential barriers, which were specically related to the guideline on anaemia. Third, participating midwives were asked to document the actual use of cut-off points for Hb and MCV in two women (pregnant women at term) they had recently seen in their practice. They were instructed to identify one woman who had received iron supplementation and one who had not received iron supplementation. In this way, we intended to check the self-reported adherence of the recommended parameters for anaemia in the guideline. The participants were asked to complete a number of pre-structured questions for each woman to document Hb value, ordering of MCV, MCV value if ordered, and decision on prescribing of iron supplementation. This information was used to assess whether the decision on prescribing of iron supplementation was consistent with the guidelines. Finally, the questionnaire comprised questions on background characteristics of the midwives: age, sex, practice (solo, duo group), urbanisation level of the practice environment (four levels), membership of professional organisation (yes/no), and professional education activities (four activities, yes/no items). The analysis of the data mainly comprised frequencies relating to the several factors that were measured. Items on the instrument on general and specic barriers were considered to reect a relevant barrier to implementation if the respondent agreed or strongly agreed to it. Comparisons between the study population and the national population were statistically tested with Chisquare tests, considering po0:05 as signicant. The case reports were recoded into dichotomous variables: practising in accordance with the recommendation of the guidelines or not practis-

Methods
A cross-sectional survey study was carried out to answer the research questions. A random sample of 260 primary-care midwives was taken from the national register of practising midwives in The Netherlands (1378 midwives in total, January 2001) (Kenens and Hingstman, 2002). We aimed to recruit at least 160 midwives, so that descriptive percentages referring to the total study population had an accuracy of at least 8% (alpha 0.05). Midwives in solo practices were over-sampled, because a lower response was anticipated in this subgroup; 60 solo practising midwives were approached. This is 23% of the sample, whereas 9% in the national population worked in solo practice. A questionnaire was posted to the sample of midwives in July 2002. Two reminders were sent. A random sample of 35 nonresponders was approached by telephone for a nonresponse analysis. A panel of three guideline developers and one guideline expert selected 14 key recommendations in the guideline in a Delphi-procedure. This procedure works well for prioritising areas of decision-making. They used the following criteria: (1) non-adherence of the recommendation would affect the essence of the guideline (yes/no); (2) non-adherence of the recommendation would affect the womens health, well-being or costs (yes/no); and (3) the recommendation is new compared with the current practice patterns in primary-care midwifery (yes/no). The 14 key recommendations are described in Table 1. The recommendations relate to the indication for ordering Hb tests, the new cut-off points in the interpretation of Hb, the use of MCV, the indication for prescribing iron supplementation and indications for referral to GPs and obstetricians.

ARTICLE IN PRESS
Guidelines on anaemia: effect on primary-care midwives in The Netherlands
Table 1 Affect of the guideline (n 160 midwives, percentages). Known Diagnosis of anaemia Healthy pregnant women: only Hb test in rst visit (i.e. at 7 12 weeks) and at 30 weeks Recommended cut-off points for normal Hb in healthy pregnant women, adjusted for haemodilution Recommended cut-off points in Negroid pregnant women (these are lower than in non-Negroid women) If low Hb: carry out MCV to distinguish between physiologically low Hb and anaemia Without anaemic complaints during the rst week post-partum: Hb test only recommended if anaemia/post-partum uxus was observed in delivery At 6 weeks after birth: only Hb test if anaemia was treated after birth Prevention and treatment Give dietary information and instructions related to iron intake No iron supplementation without further diagnosis of anaemia If low Hb: prescribe iron supplementation only if MCV value 7080 fL If anaemia is diagnosed: patient adherence is important. Alternative iron medication is not recommended After prescribing iron medication: re-test every 46 weeks Referral If Hbo6.0 [mmol/l] at rst visit: refer to GP If Hb is low and MCVo70 (fL) or X100 (fL): refer to GP During pregnancy Hbo5,6 (mmol/l): refer to obstetrician for consultation
Hb, haemoglobin; MCV, mean corpuscular volume. Midwives indicating to mostly adhere to the recommendation.

207

Agreed 118 (74%) 98 (61%) 109 (68%) 131 (82%) 117 (73%) 101 (63%) 144 (90%) 93 (58%) 86 (54%) 46 (29%) 139 (87%) 109 (68%) 115 (72%) 109 (68%)

Adhere 141 (88%) 133 (83%) 117 (73%) 130 (81%) 114 (71%) 90 (56%) 141 (88%) 118 (74%) 99 (62%) 66 (41%) 138 (86%) 109 (68%) 114 (71%) 115 (72%)

155 (97%) 157 (98%) 155 (97%) 160 (100%) 154 (96%) 150 (94%) 157 (98%) 149 (93%) 150 (94%) 112 (70%) 157 (98%) 146 (91%) 144 (90%) 149 (93%)

ing in according with the recommendation of the guideline by a midwife (rst author), using a coding system developed for this purpose that was based on the guideline. We used SPSS 10 for all analyses.

Findings
Sample and non-respondents
A total of 160 midwives responded (62%). The study population is shown in Table 2. Gendes, type of practice and urbanisation level were similar to the national population of 1378 in the year 2001 (Kenens and Hingstman, 2002). Almost all

midwives were women, and most were younger than 50 years. More than half the midwives worked in group practices. Almost all midwives were members of the national professional organisation (KNOV). Thirteen midwives participated in the nonresponse study; the remaining midwives could not be reached by telephone. All reported that they knew the guideline and had the summary card. Ten participants reported that they adhered, in most cases, to the recommended cut-off points for Hb, and 10 participants reported that they adhered, in most cases, to the recommendation to order MCV if the Hb was low. These gures were not much different in the study population, as will be shown below.

ARTICLE IN PRESS
208
Table 2 Study population (n 160 midwives). Number (%) 153 39.6 13 22 90 34 1 96 710.7 8 14 56 21 1 26 26 19 24 4 97

P. Offerhaus et al. known to 90100% of the midwives. The exception was the recommendation on patient adherence and alternative iron supplementation; this recommendation was known to 112 (70%) of the midwives. The number of midwives who reported to agreeing with, and adhering to, specic recommendations varied between 54 and 90%. The exception was, again, the recommendation on patient adherence and alternative iron supplementation; 29% agreed and 41% adhered to this recommendation. Self-reported adherence was highest (480%) for the recommendations to test Hb only at the rst appointment and at 30 weeks; for the recommended cut-off points for Hb values; for the recommendation to determine MCV values in case of a low Hb; for dietary education; and for retesting Hb every 46 weeks in women who received iron supplementation.

Characteristics Sex: female Mean age (years) Practice Solo Duo Group Locum midwife Unknown

Urbanisation Rural 42 Urbanised rural 42 Small/medium town 31 City 39 Unknown 6 155 Member of national professional organisation (KNOV) Professional education Reading national 156 midwifery journal Continuing education at 110 least twice a year Reading scientic 69 literature Participation in peer63 review group

Case reports
98 69 43 39

Being informed about the guideline

All 160 responding midwives reported knowing that the guideline existed, and 154 (96%) had the short version and the summary card of the guideline. A total of 142 midwives (82%) read the special issue of the national journal for midwives on the guideline. Most midwives reported having read the guideline very well (n 72 [45%]) or fairly well (n 53 [33%]). Few had read the book, which included scientic background information (21%). Most midwives (n 133 [83%]) had discussed the guideline with their colleagues, but few (8%) had systematically discussed it in their peer-review group. Few midwives had discussed the guidelines with obstetricians (n 40 [25%]) or GPs (n 8 [5%]).

The midwives documented their decisions on test ordering in 119 women (by 74% of the midwives) without iron supplementation and 99 women (by 62% of the midwives) with iron supplementation. In the rst cohort, the decision not to prescribe iron supplementation was appropriate in most cases (n 111 [93%]), considering the recommended pregnancy-specic cut-off points. In the second cohort, the decision to prescribe iron supplementation was appropriate in 25 cases (25%), considering the guideline recommendations. In 68 cases (69%), the Hb-values were higher than the recommended Hb cut-off points (44 cases) or no MCV value was determined, if the Hb value was lower (24 cases). In the remaining six cases, no evaluation could be carried out, because the necessary information was lacking.

Attitudes on guidelines in general

Most midwives had a positive attitude about guidelines in general. They felt that guidelines contributed to evidence-based practice and to professional development, autonomy and status of midwives (8894% agreed). They also thought that guidelines helped to reduce practice variation, enhance efciency and provide help in their daily work (8487% agreed). Few midwives thought that guidelines were too rigid or that pregnant women were too heterogeneous for guidelines (1221% agreed, two items). Most felt that guidelines protected against complaints and made the perspective of midwives explicit (6374% agreed).

Knowledge, attitude and adherence

The affect of disseminating the guideline on knowledge, attitude and self-reported adherence on the 14 key recommendations of the guideline is shown in Table 1. All but one recommendation was

ARTICLE IN PRESS
Guidelines on anaemia: effect on primary-care midwives in The Netherlands However, 101 midwives (63%) thought that guideline adherence should never be obligatory. 209

Barriers for implementation

The perceived barriers for implementing the guideline on anaemia are shown in Table 3. Most general barriers were only relevant to a few midwives. The most relevant general barriers were that GPs did not co-operate with the guideline (32%), that obstetricians did not co-operate (32%) and that specic items in the guidelines were perceived to be wrong (17%). Larger numbers of midwives mentioned barriers relating to specic aspects of the guideline. Most (59%) wanted to prescribe alternative iron supplementation or provide dietary supplements with iron. About half of the midwives (49%) found it difcult not to prescribe iron if the Hb was low but MCV was acceptable. Other barriers were a

preference to perform other tests in case the Hb was low (36%), and to test a third time, shortly before delivery (33%). They also had the idea that the lower Hb cut-off points are too low (31%), and that the period between rst visit and 30 weeks is too long (21%).

Discussion
The limitations of this study should be considered. The response rate was not different from studies in other health professionals (Gore-Felton et al., 2002). Although the non-response study did not show problems, there is still a risk of selection bias. Midwives with more favourable attitudes about the guideline may have responded in greater numbers. This means that, in the Dutch midwife population, acceptance and adherence might be lower. A discrepancy was observed between the selfreported adherence and adherence as measured

Table 3

Perceived barriers to implementing the guideline (n 160 midwives). Number (%)

General barriers General practitioners do not co-operate Obstetricians do not co-operate Specic recommendations are wrong I have not read or remembered the guideline well Difculty with changing routines I want more information Women do not co-operate Recommendations do not t with working style Lack knowledge Guideline adherence is time consuming Colleagues do not co-operate I have resistance against working according to protocols Adherence to the guideline requires nancial reimbursement I lack skills for adequate adherence to the guideline Content barriers I want to prescribe alternative iron or dietary supplements with iron Difcult not to prescribe iron if Hb is low but MCV is good If Hb is low I prefer other tests than only MCV I prefer to test Hb shorter before partus (4 30 weeks) Lower Hb values are too low, this may induce problems in partus Women respond negatively to conservative management Time period between rst visit and 30 weeks is too long I do not know how to interpret Hb in combination with MCV I am anxious to miss haemoglobinopathy Additional blood test is bothersome to the pregnant women MCV is not needed for clinical management The guideline conicts with my knowledge Diagnosis and treatment of anaemia is not my task

51 51 27 21 19 15 13 12 12 11 10 10 6 4 94 78 58 52 49 36 33 28 26 25 14 6 1

32 32 17 13 12 9 8 8 8 7 6 6 4 3 59 49 36 33 31 23 21 18 16 16 9 4 1

Hb, haemoglobin; MCV, mean corpuscular volume; *Percentages of respondents indicating agree/very agree on ve-point scale.

ARTICLE IN PRESS
210 by means of the case studies. In general, selfreported adherence gives an overestimation, and is not a valid indicator for actual practice (Adams et al., 1999). We were not able to measure actual adherence because of nancial and time constraints. The self-reports, as used in the present study, should be seen only as an indication of potential non-adherence. Women or other professionals (GPs, obstetricians) were not included in this study, although their opinions could also inuence the uptake of the guideline. The guideline on anaemia was well received by primary-care midwives in The Netherlands, and most had positive attitudes about guideline development in general. Yet, more than half of the midwives felt guidelines should not be obligatory, and one-fth thought that clients were in general too heterogeneous for guidelines. It can be stated that the dissemination strategy was successful, as nearly all midwives reported that they had received the guideline and were informed of the several key recommendations. As for the key recommendations, it can be concluded that acceptance and selfreported adherence was reasonably good. However, the case reports suggested that many women who received iron supplementation were still not diagnosed as anaemic according to recommended parameters. Other items that may require further attention in the implementation are the recommendation to test Hb after birth only if anaemia was treated, and the recommendation to enhance patient adherence and to avoid alternative iron supplementation if anaemia is diagnosed. These ndings are in line with the barriers the midwives reported in adhering to the guidelines. Furthermore, several barriers for implementation were reported, mainly relating to co-operation with GPs and obstetricians. A Swedish survey study from 1996 showed that 41% of the 134 midwives in the sample (72% response rate) adhered to the recommendations to reduce iron suplementation (Wulff and Ekstrom, 2003). Interestingly, 79% of senior obstetricians and 28% of pregnant women reported adhering to this recommendation. We can only speculate about the explanation of these ndings, but they might suggest that midwives were more informed of, and receptive to, clinical research evidence than pregnant women, but less than obstetricians. The implementation of guidelines into health care is widely recognised as a complex process. The process can be divided into different various stages, such as dissemination, implementation and continuation. The transition from one stage to the other can be affected by various barriers, and is unlikely to be a linear process (Fleuren et al., P. Offerhaus et al. 2004; Grol and Wensing, 2004;). For example, a health professional may be prepared to use the guideline but, in daily practice, women may not cooperate. In this case, the health professional may come back on the intention to implement the guideline in her professional behaviour. It is essential to identify barriers for change in each stage and to accommodate these in the implementation strategy. The ndings of this study showed that the dissemination stage was passed successfully, as nearly all midwives had the short version and summary card of the guideline. The selfreported adherence to the recommendations in the guideline was lower, which was related to specic barriers. Important barriers were a tendency to prescribe iron supplementation in the case of low Hb values, and a tendency to test Hb more often than recommended and to use tests other than MCV. Convincing evidence to support these recommendations may be sufcient to change behaviour in some midwives, but additional activities are probably needed to change the routines of others. Interactive education, such as outreach visits or quality circles, perhaps using opinion leaders, may be needed to change professional routines of midwives. The practice implication is that the recommendations, of which adherence was low, (i.e. lower than 70%), should be addressed in the implementation activities by the national organisation of midwives. A growing number of midwives in The Netherlands participate in small peer-review groups. Research showed that these peer-review groups might have a positive effect on changing clinical practice (Engels et al., 2003). Therefore, a policy recommendation is to discuss the recommendations in peer-review groups, to analyse impeding and facilitating factors in adherence, and to design an appropriate implementation strategy that is adapted to these factors. This also gives the opportunity to adapt the implementation strategies to the local circumstances, such as the co-operation with local GPs or obstetricians. In addition to the above mentioned analyses of impeding and enhancing factors, further research should examine the effectiveness of the specic interventions that were used to improve current practice.

References
Adams, A.S., Soumerai, S.B., Lomas, J., et al., 1999. Evidence of self-report bias in assessing adherence to guidelines. International Journal for Quality in Health Care 11, 187192.

ARTICLE IN PRESS
Guidelines on anaemia: effect on primary-care midwives in The Netherlands
Amelink-Verburg, M.P., Daemers, D.O.A., Rijnders, M.E.B., 2000. Anaemia in primary care midwives practice [Anemie in de eerstelijns verloskundige praktijk]. Bilthoven, Leiden. Burgers, J.S., Cluzeau, F.A., Hanna, S.E., et al., 2003. Characteristics of high-quality guidelines: evaluation of 86 clinical guidelines developed in ten European countries and Canada. International Journal of Technology Assessment in Health Care 19, 148157. Engels, Y., Verheijen, N., Fleuren, M., et al., 2003. The effect of small peer group continuous quality improvement on clinical practice of midwives in the Netherlands. Midwifery 19, 250258. Field, M.J., Lohr, K.N. (Eds.), 1990. Clinical practice guidelines: direction for a new program. National Academy Press, Washington DC. Fleuren, M.A.H., Wiefferink, C.H., Paulussen, T.G.W.M., 2004. Determinants of innovation within health care organizations: Literature review and Delphi study. International Journal for Quality in Health Care 16, 107123. Gore-Felton, C., Koopman, C., Bridges, E., Thoresen, C., Spiegel, D., 2002. An example of maximizing survey return rates. Methodological issues for health professionals. Evaluation And The Health Professions 25, 152168.

211

Grol, R., Grimshaw, J., 2003. From best evidence to best practice: effective implementation of change in patients care. The Lancet 362, 12251230. Grol, R., Wensing, M., 2004. What drives change: barriers and incentives to achieving evidence based practice. The Medical Journal of Australia 180, 5760. Kenens, R., Hingstman, L., 2002. Figures from the registration of midwives. Peiling 2001. NIVEL, Utrecht. Looker, A.C., Dallman, P.R., Carroll, M.D., et al., 1997. Prevalence of iron deciency in the United States. The Journal of the American Medical Association 277, 973976. Peters, M.A.J., Harmsen, M., Laurant, M.G.H., Wensing, M., 2003. Room for improvement? Barriers and opportunities for improving patient care. WOK, Nijmegen. Wildschut, H.I.J., Hoope-Bender ten, P., Borkent-Polet, M., et al., 1998. Audit of pre-natal screening in primary and secondary care. Tijdschrift voor Verloskundigen 23, 764769. Wulff, M., Ekstrom, E.C., 2003. Iron supplementation during pregnancy in Sweden: to what extent is the national recommendation followed? Acta Obstetricia et Gynecologia Scandinavica 82, 628635.