SEARCH

SEARCH BY CITATION

Keywords:

  • Birthrate+;
  • labor;
  • personnel staffing and scheduling;
  • quality improvement

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References
  9. Supporting Information

Please cite this paper as: Allen M, Thornton S. Providing one-to-one care in labour. Analysis of ‘BirthratePlus’ labour ward staffing in real and simulated labour ward environments. BJOG 2012; DOI: 10.1111/j.1471-0528.2012.03483.x.

Objective  To assess the ability of the ‘Birthrate Plus’ (BR+) midwife staffing system to cope with variability of workload on labour wards.

Design  Retrospective analysis of labour ward workload and computer simulation of labour wards.

Setting  The labour ward of a city hospital.

Population  A total of 5800 births (1 year).

Methods  The variation in births by time and day was analysed over a 1-year period. Three months of BR+ data were analysed for variation of workload by case mix. A computer simulation model was built to allow prediction of the impact of changing resource levels or shift patterns, and to forecast the impact of changing number of births per year.

Main outcome measures  Labour ward overloading (when either the number of women or the BR+ Workload Index exceeds the scheduled midwife availability).

Results  Labour ward overload occurred 37% of the time when applying the BR+ method. Underlying patterns of workload were present and simulation indicated that overload could be reduced by 15–25% if available resources were matched more closely to known patterns of workload. Simulation also indicated that smaller units are predicted to suffer from overload more often than larger units, and are more prone to severe overload.

Conclusions  The BR+ formula for midwife staffing leaves labour wards vulnerable to significant periods of overload. Matching resource levels to known patterns of workload may reduce the occurrence of overload. Simulation indicates that smaller units need higher relative staffing levels to provide the same level of 1:1 care to mothers in labour.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References
  9. Supporting Information

There is limited information on the impact of midwife staffing on birth outcomes,1 but shortfalls of midwives on labour wards have been associated with a higher incidence of ‘near misses’ and adverse events such as delays in time-critical procedures.2 Tucker et al.3 also reported an association between lower midwife levels and an increased incidence of neonatal resuscitation. In a recent Cochrane review and meta-analysis4 the authors identified evidence that continuous support in labour led to a higher probability of a vaginal birth, a lower use of anaesthetics and lower probabilities of instrumental delivery, caesarean section or low 5-minute Apgar scores. These reductions in likelihood of interventions that accompany continuous care in labour are likely to have cost benefits to the provider. Additionally, lack of continuous perinatal support is associated with poor experience for the mother; in the UK a recent survey by the Care Quality Commission5 found that 25% of responders reported being left alone by midwives or doctors at a time when it worried them (15% were left alone during labour, 6% shortly after birth and 5% during labour and shortly after birth). There is therefore a gap between the standard of care desired by the mothers, and the standard of care experienced or perceived by the mothers. It is now recommended by the UK Department of Health and by UK medical colleges (Royal College of Anaesthetists, Royal College of Midwives, Royal College of Obstetricians and Gynaecologists and the Royal College of Paediatrics and Child Health) that women should receive dedicated one-to-one care during labour and in the immediate time after childbirth.6–9 However, the lack of empirical data linking any particular midwife:mother ratio to outcome, beyond the benefits of continuous care being demonstrated, should be noted.

A challenge facing care-providers is that the required resources may vary depending on demand and case mix. In emergency departments it has been reported that shortfalls in staffing may often be a result of variability in demand rather than an inability to meet average demand.10 In the UK, the Royal Colleges6,9 have endorsed the Birthrate Plus (BR+)11 system as a planning tool to predict the number of midwives required on a labour ward because it takes demand and case mix (acuity) into account. By 2003, 101 maternity services across 117 sites in the UK had adopted BR+ for midwife staff planning,12 and BR+ is still the method endorsed by the Royal College of Midwives.9 The tool is now also used in other countries, for example a project to test its usefulness in Australia has been in progress in New South Wales for the past 3 years.1 Given the reported gap between desired and perceived level of care5 we sought to analyse data obtained during a BR+ assessment aided by computer simulation to model the level of care provided under BR+ guidelines. Though BR+ is the accepted standard for setting midwife levels there is little published assessment of the method, and the King’s Fund has recently called for objective research on the effectiveness of BR+.1 Our response to this call is this first objective study examining how well BR+ performs in its ability to provide one-to-one care during labour.

We looked at variation in workload in maternity units, how much reserve resource (required to cope when workload is above average) is needed to cope with these variations (BR+ adds a 15% reserve), and whether there were any underlying predictable patterns of demand. One statement made in the BR+ method is ‘As clients are admitted and are in labour at any time of day and night, then equal numbers of midwives are needed on each shift’ (see ref. 11, p.101). Emergency department planning has used manufacturing techniques in its analysis of supply and demand13 and in this study we have sought to bring the same techniques to bear on examining supply (of midwifery staff) and demand in the labour ward.

Computer simulation allows the performance of experiments that are very difficult to carry out in reality. One aspect of the labour ward (such as number of deliveries per year, or number of midwives on each shift) can be changed while everything else is held constant to forecast the expected change in performance. Simulation has been widely used to model and analyse healthcare environments,14,15 including its use to predict required nurse staffing.16,17 We used computer simulation of labour wards to investigate the expected effect of changing midwife numbers throughout the day and week. We also investigated how size of units affects the unit’s ability to cope with variation in workload.

The objectives of this study were (1) to provide the first empirical objective evidence for how well the BR+ midwife staffing method supports the provision of one-to-one care during labour, and (2) to investigate the impact of the size of maternity units on the ability to provide one-to-one care during labour.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References
  9. Supporting Information

We examined historical data from a labour ward and compared the numbers of women on the ward and the total workload (‘Workload Index’ as assessed by BR+, which takes patient acuity into account) with the recommended level of midwife support as calculated using BR+. We used computer simulation to investigate the potential of alternative staffing schedules and to investigate how changing births per year affects the ability to provide one-to-one care during labour.

Data sources

Two data sources were used for this study from the University Hospital Coventry, which is a city hospital providing support for ∼6000 births per year. No patient identifiable information was used in this study.

Hospital record of births for 1 year: These records were used to examine the daily, weekly and yearly pattern of deliveries (categorised by type of delivery). The records contained time of birth and type of birth, but not details on length of stay in labour ward or interventions that are used to calculate the BR+ Category.

Birthrate Plus data for 3 months from the same hospital: These records were collected previously for calculation of staffing levels using the BR+ formula. These data included time of arrival and departure and information on a range of key characteristics of the labour and birth (these characteristics are used to classify labours into five distinct BR+ categories).

BR+ calculations

Birthrate Plus is used as a guide to the number of midwives required on a labour ward. BR+ calculates the number of midwives required by adding up the total time mothers spend on the labour ward and multiplying each of five categories of mother (according to interventions received during labour) by a multiplier, allowing for increased midwife support for higher acuity labours.11 The end result is the ‘Workload Index’ which describes the total midwife staffing recommended for coping with the caseload mix.

Using the BR+ data we reconstructed the number of mothers on the ward, along with their BR+ Category, over the 3 months and calculated (1) the percentage of time when there were more mothers than midwives present (as recommended by BR+) on the labour ward, and (2) the percentage of time that the BR+ Workload Index in any given hour exceeded the number of midwives present for that hour. More details on BR+ are available in the Supplementary material, Appendix S1.

Model description

The simulation model mimicked the arrival pattern of mothers (separating elective and spontaneous births). Mothers would then be assigned to a BR+ Category according to the distribution observed in the hospital and their length of stay on the labour ward would be sampled from a distribution for their particular BR+ Category. The model therefore replicates arrival rates and patterns and length of stay patterns. Once validated against the current state (using actual BR+ data to reconstruct the number of mothers present and the Workload Index across 3 months) the model was used to test the likely outcomes of alternative scenarios. More details on the simulation model are available in the Supplementary material, Appendix S1.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References
  9. Supporting Information

Analysis of patterns in the 1-year births data set

The average number of births per day over the year was 16.2 with a standard deviation (SD) of 4.0. On 80% of days there were 11–22 deliveries, whereas 10% of days had fewer than 11 deliveries and 10% had more than 22 deliveries.

The number of births per day was analysed by month. Over 12 months, the average number of births per day did not significantly change by month (P > 0.05, analysis of variance [anova]). The number of births by day of week varied significantly (P < 0.001, anova), with average daily deliveries being 20% higher on weekdays than at weekends. If caesarean sections were removed there was no effect of day of week on the number of births (P > 0.05, anova).

The number of births varied by hour of day (P < 0.001, anova) with a marked pattern caused by elective caesarean section deliveries, which tended to occur between 09.00 and 12.00 hour on weekdays. During this 3-hour period the number of births was 60% higher than the average of the rest of the day. When elective caesarean section deliveries were removed, the variation in the number of births by hour of day remained (P < 0.01, anova) but the magnitude was smaller (Figure 1). On weekdays, the impact of time of day made a maximum difference of 0.11 non-elective births per hour (against an average of 0.70) whereas the elective caesarean section deliveries could change delivery rates by up to 0.67 births per hour depending on time of day.

image

Figure 1.  Births throughout the day on weekdays. The solid bars show all births except for elective caesarean sections and the open bars show elective caesarean sections.

Download figure to PowerPoint

Analysis of patterns in 3 months BR+ data

The proportion of women in each BR+ Category and their length of stay on the labour ward are shown in Table 1. Using the BR+ formula a total midwife resourcing allocation of 45 whole time equivalents was calculated, which (after allowing for the administrative and sickness/holiday time) gives an average of 8.2 midwives on the labour ward at any one time. To calculate how often the number of mothers, or the Workload Index, exceeds the number of midwives, we have used an estimate of eight midwives being present on the labour ward at any one time (as in reality the number of midwives on a shift will be rounded to a whole midwife).

Table 1. Three month birth data showing frequency of BR+ categories, average length of stay on the labour ward, SD of length of stay and % CV (SD/mean as %)
BR+ category*% of birthsLOS on labour ward (minutes)% CV
AverageSD for LOS 
  1. CS, caesarean section; LOS, length of stay; CV, coefficient of variation.

  2. Category 1–5: labours categorised according to requirements for interventions; Category 1: no intervention, to Category 5: highest level of intervention (More detail is available in Supplementary material, see Appendix S1).

Category 11833317151
Category 22442322954
Category 31760028648
Category 4 (excluding elective CS)1675237149
Category 4 (elective CS only)1017717738
Category 516100383683

Using recorded data on arrival time and length of stay we reconstructed the number of women (and their total Workload Index) on the labour ward over time. The mean number of women present on the labour ward was 5.9 (SD 2.5, median 6) with an average Workload Index of 7.4 (SD 3.1, median 7.2). The mean ratio of midwives to mothers was 1.3 (median 1.3) and the mean ratio of midwives to Workload Index was 1.1 (median 1.1).

For 36% of the time there was a greater Workload Index (which takes into account greater midwife use for more complex cases) than the number of midwives available and for 13% of the time there were more mothers than midwives. The maximum number of mothers on the labour ward during the 3-month period was 15, with a maximum Workload Index of 18.9 (18.9 midwife hours required in 1 hour according to the BR+ formula).

There was a clear pattern in the workload on the labour ward (Figure 2); the average number of women on the labour ward increased during the day on weekdays (being 30–40% higher than nights and weekends), peaking at about midday. It was reduced overnight and at weekends. A similar pattern was seen in the occurrence of work overload, during nights and weekends the number of mothers exceeded the number of midwives 5–10% of the time but this increased to 25–30% during the day on weekdays. During the peak period (09.00 to 13.00 hour weekdays) the hourly BR+ total Workload Index exceeded the allocated number of midwives ∼65% of the time.

image

Figure 2.  Patterns in the number of mothers being present on the labour ward (left panel), the percentage of time when the number of mothers exceeds the BR+ midwife allocation (centre panel) and the percentage of time when the BR+ Workload Index exceeds the BR+ midwife allocation (right panel). Data are taken from 3 months of BR+ data and are separated into weekdays (solid circles) and weekends (open circles).

Download figure to PowerPoint

Validation of simulation model

The output of the simulation model was compared with the actual 3-month BR+ data. In all key indicators the model produced results that were within 5% of the actual data. The model produced an average labour ward occupancy of 5.9 women (cf. observed 5.7) and average hourly BR+ Workload Index of 7.4 (cf. observed 7.1). The number of women exceeded the number of midwives 13% of the time in the model (cf. observed 14%), and the hourly BR+ Workload Index exceeded the number of midwives 36% of the time (cf. observed 37%). The same patterns of fluctuations in workload and overload were also observed. This gave the required confidence to use the model to explore alternative scenarios.

Relationship between staffing levels and incidence of overload

We simulated a labour ward delivering 6000 babies per year with varying midwife levels (Figure 3). As the number of midwives was increased the probability of overload reduced. To guarantee that there were more midwives than mothers 95% of the time, the average midwife:mother ratio on the ward needed to be ∼1.8 (standard BR+ calculations set this ratio at about 1.4). If Workload Index is taken as a more reliable guide to workload on the labour ward then to assure that there is sufficient resource to cover Workload Index 95% of the time, the average midwife:mother ratio on the ward needed to be ∼2.2, significantly higher than BR+ guidelines.

image

Figure 3.  Effect of altering the number of midwives on duty on the probability that the labour ward may become overloaded. Labour ward overload was measured either by the number of mothers on the labour ward exceeding the number of midwives (solid circles) or the Workload Index exceeding the number of midwives (open circles). Results come from a simulated labour ward with 6000 births per year (average number of mothers on labour ward 6.2, average Workload Index 7.8).

Download figure to PowerPoint

Probability of labour ward overload was significantly higher during the day on weekdays. The model forecast that performance of the unit could therefore be improved by increasing resources at just the time of these predictable increases in load; a 25% reduction in occurrence of overload could be achieved with only a 4% increase in budget. Alternatively a no-cost option could be used and staffing levels could be reduced on Saturday night and all of Sunday (when demand is lowest and cost per midwife is highest) and re-applied at peak load during weekdays. In this no-cost option a 15% reduction in occurrence of overload could be achieved.

Effect of size of unit on probability of labour ward overload

We performed simulations for labour units of different sizes, delivering 1000–12 000 births per year (all with the same mix of patients as described in Table 1), while keeping the ratio of midwives to deliveries constant. The model forecasts the variation in mothers present (along with their BR+ Category) as the number of mothers arriving each day changes. An audit was run every hour in this virtual labour ward to count the number of mothers present with the hourly Workload Index for comparison against the number of midwives present as recommended using the BR+ formula. As the size of unit increased the amount of time that the labour ward was overloaded reduced (Figure 4). Using the BR+ calculation small units (∼2000 births per year) were forecast to have more mothers than midwives 16% of the time whereas the larger units (∼8000 births per year) were overloaded 10% of the time. The percentage of time that the hourly Workload Index exceeded the number of available midwives similarly fell from ∼45 to ∼30%.

image

Figure 4.  Effect of the size of unit on the number of whole time equivalent midwives required to provide one-to-one cover in labour. The solid circles represent the number of midwives needed per 1000 annual births to provide one-to-one cover 95% of the time. The open circles represent the number of midwives needed per 1000 annual births to cover the Workload Index 95% of the time.

Download figure to PowerPoint

The severity of potential overload was significantly worse for smaller units. In a unit delivering 2000 births per year the Workload Index could rise to twice the number of allocated midwives (occurring 6% of the time) whereas this level of severe overload was very rare in a unit delivering 8000 births per year (occurring just 0.1% of the time).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References
  9. Supporting Information

A problem facing any clinical system where women may turn up at random is the challenge of fluctuating workload. Clinical systems may frequently become overloaded, with associated risk to women, if resource levels are set to average demand levels rather than taking the range of expected workload into account.18 Though the pattern of workload may appear random, the range of normal fluctuation is measurable and predictable. Given a fixed number of midwives it is possible to forecast the proportion of time that the unit has at least as many mothers as midwives, or how much of the time the BR+ Workload Index is covered by the planned staffing. The BR+ formula for calculating number of midwives allows for 15% extra resource (above the average Workload Index) for coping with fluctuation in workload. We found that in practice the Workload Index exceeded planned resource 36% of the time and the number of mothers exceeded the number of midwives 13% of the time. These periods of labour ward overload are likely to be when the quality of service is most significantly affected and when risk to the mothers and babies is increased.6 It is also difficult to retain overworked midwives.19 An analytical approach to labour ward demand allows for more informed decisions on how staffing levels impact overload.

We identified a clear underlying pattern of increased workload during weekdays, which was associated with scheduling of elective caesarean sections on weekdays. BR+ recommends that the numbers of midwifes on the labour ward be kept constant throughout the day and week, but many units may have predictable patterns.20 Walley et al.21 have recommended that resource planning in the emergency department should aim to predict and prevent overloading rather than reacting to overloading that has already developed. We similarly recommend that workload patterns are analysed when deciding on resource levels across different shifts by weekday and by time of day. Modelling showed that that this approach would reduce the number of occasions when there were insufficient midwives to provide one-to-one care.

Data showed that workload on the labour ward can change very rapidly. Matching staffing levels to known demand patterns is a good start, but there is still unpredictability. Our results show that the limited (15%) reserve for variation allowed in the BR+ formula is insufficient to cope with workload fluctuations. A target reserve of 15% is common in healthcare systems, since Bagust et al.22 recommended 85% as the target bed occupancy in a 200-bed hospital, but smaller systems such as a labour ward experience greater variability in workload and require a greater reserve than a large system such as bed occupancy in a hospital. Labour wards behave more like emergency departments than a 200-bed hospital. Bringing in additional midwives during unexpected peak demand may help, but only if the speed of response is timely, rather than the midwives arriving just as the peak in workload is subsiding by natural variation. If a higher constant level of staffing is used, thought should be given to how midwives use their time when demand is lower—are there non-time-critical tasks now being performed by other staff that could be performed by midwives during quieter periods? The increased costs of a higher number of qualified staff may be offset by the reduction in interventions that accompanies continuous care in labour.4 Some units may consider resourcing elective caesarean sections separately (as they represent a planned activity that only occurs at certain times). This is certainly feasible, although such ‘carve-out’ carries a risk that midwives dedicated to one work-stream may find themselves idle while midwives allocated to another work-stream are simultaneously overworked. Use of auxiliary staff may also help to cope with peaks in demand, though again carve-out of different types of work for different staff can limit flexibility of moving available resources to current demand.

We used the model to examine the effect of unit size on overloading. BR+ calculations scale the recommended resource level in direct proportion to the number of births per year. Smaller units however have a larger relative variability in their workload compared with larger units, making them more prone to periods of excessively high or low workloads. When we took one unit with a given mix of case loads and varied the number of births in the simulation model, we found that units with 1000–2000 births were overloaded twice as often as a unit of 8000 births per year. The magnitude of overloading was also significantly worse for smaller units where Workload Index could rise to twice the level of planned resource. These periods of severe overload were much rarer in a unit delivering 8000 births per year. Smaller units therefore require a proportionally increased number of midwives (or the ability to rapidly bring in additional resources when required) to maintain a low risk of overloading.

Generalisation of results

Our analysis and model have been based on one particular hospital (though we have used simulation to extend the study to other ‘virtual’ hospitals). Simulation models may be adjusted to local circumstances (such as case mix) but the key findings may be generalised without more bespoke simulation. The BR+ formula is not a guarantee of providing one-to-one cover in labour, and there may be significant periods (a third of the time in this study) when the Workload Index, which takes patient acuity into account, is in excess of the midwife resource available. Many hospitals may also have predictable fluctuations in demand and we would recommend that those using BR+ look for, and adapt midwife numbers to, these predictable daily and weekly patterns. The simulation predicts that larger units will experience less variation in workload than smaller units. Changing case mix may lead to different absolute numbers of midwives required, but the principle that larger units help to reduce the range of variation and avoid periods of excessive overload will be maintained.

Limitations

This analysis deals with an isolated labour ward. Midwife staffing levels for other work are not included (e.g. midwives involved in transfers of mothers between units, midwives undertaking supervising activities, or midwives dealing with women not in established labour and not admitted onto the labour ward). We have assumed, as with published guidelines,6 that auxiliary staff are additional to, and not a replacement for, midwives. Though midwives should be present to provide continuity of care even when the mother is also under the care of other medical practitioners,6 this paper has not sought to draw any conclusions on the required number of other clinical staff (such as anaesthetists and obstetricians). We have not attempted to make any extrapolation to clinical outcome in this model.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References
  9. Supporting Information

An analysis of variation of labour ward workload allows prediction of the percentage of time that the number of allocated midwives is likely to be insufficient to provide one-to-one care or to cover the Workload Index. While BR+ sufficiently covers average workload we found that the BR+ allocation of midwives could be overloaded 14% of the time when simply counting mothers on the labour ward, or 36% of the time when allowing for greater resource demand of more complex cases. In addition to understanding the extent of expected variation of workload, a predictable daily and weekly workload pattern was identified. Rather than allocating constant midwife resources across the day and week, an understanding of such patterns allows for a closer match of planned resources to expected workload. The size of a unit is also predicted to affect the likelihood of overload, with smaller units significantly more susceptible to work overload, and the severity of the overload significantly worse than in larger units.

Disclosure of interests

None relevant.

Contribution to authorship

MA undertook all data collection, analysis and computer simulation, and wrote the paper. ST provided expert knowledge of the labour ward environment, helped to shape the research question, checked the analysis against real-life experience, checked that the computer model reflected real-life practice and helped to write the paper.

Details of ethics approval

No patient-identifiable information was used in this study. The hospital Research & Development department was approached and advised that as no patient-identifiable information was being used no formal ethical approval was required.

Funding

MA is funded by the National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care (CLAHRC) for the South West Peninsula. The views expressed in this publication are those of the authors and not necessarily those of the National Health Service, the NIHR or the Department of Health, United Kingdom.

Acknowledgements

Thanks go to the midwives of University Hospital Coventry, especially the Modern Matron Claire Allan, who helped to provide background information for shaping the project.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References
  9. Supporting Information
  • 1
    Sandall J, Homer C, Sadler E, Rudisill C, Bourgeault I, Bewley S, et al.Staffing in Meternity Units (Getting the Right People in the Right Place at the Right Time). London: The King’s Fund, 2011.
  • 2
    Ashcroft B, Elstein M, Boreham N, Holm S. Prospective semistructured observational study to identify risk attributable to staff deployment, training, and updating opportunities for midwives. BMJ2003;327:584A586A.
  • 3
    Tucker J, Parry G, Penney G, Page M, Hundley V. Is midwife workload associated with quality of process of care (continuous electronic fetal monitoring [CEFM]) and neonatal outcome indicators? A prospective study in consultant-led labour wards in Scotland. Paediatr Perinat Epidemiol2003;17:36977.
  • 4
    Hodnett ED, Gates S, Hofmeyr GJ, Sakala C, Weston J. Continuous support for women during childbirth. Cochrane Database System Rev2011;2:CD003766.
  • 5
    Care Quality Commission. Women’s experiences of maternity services 2010. [http://www.cqc.org.uk/aboutcqc/howwedoit/involvingpeoplewhouseservices/patientsurveys/maternityservices.cfm]. Accessed 1 December 2010.
  • 6
    Royal College of Anaesthetists, Royal College of Midwives, Royal College of Obstetricians and Gynaecologists, Health RCoPaC. Safer Childbirth: Minimum Standards for the Organisation and Delivery of Care in Labour. London: RCOG Press, 2007.
  • 7
    National Collaborating Centre for Women’s and Children’s Health. Intrapartum Care: Management and Delivery of Care to Women in Labour (CG55). In: National Institute for Health and Clinical Excellence. London: RCOG Press, 2007.
  • 8
    Department of Health. Maternity Standard, National Service Framework for Children, Young People and Maternity Services. London: DH Publications, 2004.
  • 9
    Royal College of Midwives. Guidance Paper on Staffing Standard in Midwifery Services. London: Royal College of Midwives, 2009.
  • 10
    Higginson I, Whyatt J, Silvester K. Demand and capacity planning in the emergency department: how to do it. Emerg Med J2011;28:12835.
  • 11
    Ball JA, Washbrook M. Birthrate Plus: A Framework for Workforce Planning and Decision Making for Midwifery Services. Books for Midwives Press, Hale, Cheshire, UK: Hochland & Holland, 1996.
  • 12
    Ball J, Bennett B, Washbrook M, Webster F. Birthrate Plus Programme: a basis for staffing standards?Br J Midwifery2003;11:2646.
  • 13
    Walley P. Designing the accident and emergency system: lessons from manufacturing. Emerg Med J2003;20:12630.
  • 14
    Mustafee N, Taylor SJE, Katsaliaki K, Brailsford S. Distributed simulation with COTS simulation packages: a case study in health care supply chain simulation. In: Proceedings of the 2006 Winter Simulation Conference. Monterey, CA, USA. December 3–6, 2006; pp. 113642. Download from Brunel University Research Archive, 2006 (http://dspace.brunel.ac.uk/bitstream/2438/4001/3/Dist%20Sim%20with%20Simul8%20HLA.pdf).
  • 15
    Katsaliaki K, Mustafee N. Applications of simulation within the healthcare context. J. Operational Research Society2011;62:143151.
  • 16
    Duraiswamy N, Welton R, Reisman A. Using computer simulation to predict ICU staffing needs. J Nursing Admin1981;11:3944.
  • 17
    Griffiths JD, Orcie-Loyd N, Smithies M, Williams JE. Modelling the requirement for supplementary nurses in an intensive care unit. J Operat Res Soc2005;56:1235.
  • 18
    Litvak E, Buerhaus PI, Davidoff F, Long MC, McManus ML. Managing unnecessary variability in patient demand to reduce nursing stress and improve patient safety. Jt Comm J Qual Patient Saf2005;31:3308.
  • 19
    Curtis P, Ball L, Kirkham M. Why do midwives leave? (Not) being the kind of midwife you want to be. Br J Midwifery2006;14:2731.
  • 20
    MacFarlane A. Variations in number of births and perinatal mortality by day of week in England and Wales. BMJ1978;2:16703.
  • 21
    Walley P, Silvester K, Steyn R. Managing variation in demand: lessons from the UK National Health Service. J Healthc Manag2006;51:30922.
  • 22
    Bagust A, Place M, Posnett JW. Dynamics of bed use in accommodating emergency admissions: stochastic simulation model. BMJ1999;319:1558.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References
  9. Supporting Information

Appendix S1. Further information of BR+ calculations and model description.

FilenameFormatSizeDescription
bjo3483_sm_AppendixS1.pdf49KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.