A recent study co-authored by Professor Graham Cookson from our research team and published in the Health Service & Delivery Research Journal focused on analysing the role of the workforce in the delivery of maternity services. After estimating several models they concluded that staffing levels may have either positive or negative effects, depending on the outcome under examination. However, they demonstrated that the role of doctors and midwives is more important in higher-risk settings and they underlined that providers should be cautious with the use of support workers in such cases.
While this study utilized data for the NHS from 2010/11, we plan to extend this work using longitudinal data at the trust level and linking them with individual maternity records from the Hospital Episodes Statistics (HES) database, covering the period from 2004/05 to 2012/13. We believe this will improve our knowledge regarding the implications of staffing levels and skill mix on maternity outcomes since variation at the trust level, the role (if any) of dynamics and any omitted variable problems will be more properly accounted for.
Here we present some early descriptive trends regarding staffing levels and skill mix variables. In order to do so, we merged two distinct datasets. The first one contains information about medical and clinical staff in Obstetrics and Gynaecology for the years 2004 to 2014 and it is provided by the Health & Social Care Information Centre (HSCIC) Workforce Census. It contains information regarding the total full time equivalent (FTE) by trust and grade as at the last day of September of each year. We also used the NHS Maternity Statistics which are collected using HES records and provide us with details on every delivery taking place in all NHS hospitals. These data were available from the HSCIC website for the period between 2004/05 and 2012/13. Regarding the Maternity Workforce Census data, combining the various years into a single dataset was quite straightforward, despite the fact that some grades (especially for the clinical staff, i.e. nursing assistant practitioners) were recorded only from 2011/12 onwards. Yet, harmonizing the various years of the NHS Maternity Statistics was more painful since the variable names, the unit configuration and the spreadsheet format was changing from year to year (although this issue will be discussed in detail in a future post).
Using the trust code and the year as identifiers the two datasets were merged to construct a single database containing longitudinal information on both the maternity workforce and delivery outcomes at the trust level. In order to explore some trends regarding the staffing levels and the skill mix in the NHS maternity services, the following variables were constructed:
1. FTE doctors per 100 maternities
2. FTE consultants per 100 maternities
3. FTE midwives per 100 maternities
4. FTE support workers per 100 maternities
5. FTE healthcare assistants per 100 maternities
6. FTE all staff per 100 maternities
7. FTE managers per 100 maternities
8. Doctors to midwives ratio
9. Consultants to doctors ratio
10. Support workers to midwives ratio
11. Healthcare assistants to midwives ratio
12. Managers to total staff ratio
To take into consideration the variation between regions, the combined dataset was collapsed by year and Strategic Health Authority (SHA) and each index is plotted separately for each one of the ten SHAs as well as for the country as a whole. The next graph displays the evolution of the FTE doctors per 100 maternities. From 0.69 in 2004/05 it has increased to 0.76 in 2012/13, yet there is considerable variation at the regional level. Notably, it has decreased, on average, for trusts located in the South West and the South East Central SHAs.