Babies are more likely to die at birth or soon after if the midwife is in her first year on the job, Otago University researchers say.
The contentious findings have been published today in the International Journal of Gynaecology and Obstetrics.
Lead author Beverley Lawton said researchers from the university's Wellington-based Women's Health Research Centre and Illinois University matched the level of experience of New Zealand midwives to data on baby mortality in New Zealand between 2005 and 2009.
She said they found a higher likelihood of a baby dying at or near birth if the midwife - with midwifery-only training - was in the first year out of midwifery-only training, compared to midwives with between five and nine years' training.
"If the midwife is trained under one year from qualification as a lead maternity carer the babies had a perinatal mortality rate of 8.1 percent per 1000 births, and if they were 5 to 9 years qualified, it was 6.1 percent per thousand births. So there's a difference for the more experienced midwives of 2 less deaths per thousand."
Dr Lawton said similar findings were not found for nurse-trained midwives.
She said they also found a statistically significant trend relating to the midwifery-only trained midwives in their first year.
"So it wasn't just a one-off thing and as they got more experience, there was a trend to less deaths."
Dr Lawton said the government should review New Zealand's maternity training system urgently.
"I think it should shine the spotlight particularly on supervision and that first year. And I think we should be looking to review what's happening, and quite urgently. At a minimum we should be offering supervision for our maternities who are out there now being looked after by a first year."
Dr Lawton said there was no requirement in New Zealand for supervision in the first years after graduation for midwife-only midwives, although a mentorship programme was introduced in 2011. "Due to well publicised deaths of babies, the mentorship programme has been strengthened but the researchers believe that until this is changed to supervision this is unlikely to make a substantial difference."
Dr Lawton said it was fair to look at outcomes for lead maternity carers, because they were responsible under the Health Ministry's section 88 maternity rules for understanding the risks and "navigating that mother and her pregnancy into the right care".
But the chair of the Midwifery Council, Judith McAra Couper, said the research was six to 10 years old, and since then the Bachelor of Midwifery qualification had gone from three to four years.
"And the last year is now 80 percent of clinical, which is in effect an intern year, and the graduates go straight from that into a mandatory midwifery first year of practice programme where they receive mentoring, clinical support, further education and professional development."
Dr Couper said it was hard to see how it could be improved. "We surround our midwives with a fantastic safety net of their obstetric colleagues, the hospital midwives, their practice partners, their mentor. I'm not sure what else we could add in here."
She added the safety of mothers and babies was always the council's top priority. "While we will examine this research at our next Council meeting, we can assure women and their families that midwives are providing a high quality and safe service."
A midwifery adviser for the College of Midwives, Lesley Dixon, also said major changes had been made to maternity care since 2009. "There's been a huge number of changes during that time to the undergraduate programme, to the continuing supervision of our graduates, and we know that perinatal mortality rates have been reducing."
She said the research will worry women for no reason. "What this does is it's scaremongering for women. It actually makes women worried about their care provider and that's a concern essentially when we have worked really hard to ensure that women have a safe and high quality maternity system."
The Health Ministry was not available to be interviewed, but said in a statement it too has questions about the methods used in the study.
It said further research would be needed to support the study's conclusions, and any further research should use current data.
Health Minister Jonathan Coleman said he would take the findings seriously, but needed to look more closely at the study first. "We've got to dig into this and find out what really went on, but initial advice is there's concerns about the methodology in the study, which covers the period 2005 to 2009. So look I'm taking it very seriously but I've got to get some more advice about it."