Earlier this week I discussed why differences in definitions of live and still birth make comparisons of international statistics on infant mortality misleading. But these divergences in reporting requirements are not the only reason that international comparisons of morality rates are unreliable and inaccurate.
In other cases, bureaucracy intrudes so that even those who meet the definition of live birth may not end up in statistics. In France, in order for the baby to be given a birth certificate as a live birth, one must have “a medical certificate [that] attests that the child was born ‘alive and viable’.” On the other hand, “[i]n the absence of a medical certificate attesting that the child was born ‘alive and viable,’ the civil service officer only establishes a certificate of a child without life,” in lieu of a birth certificate. Overall, “this procedure applies, on one hand, to children born alive but not viable, and on the other, to children stillborn after a term of 22 weeks…or with a weight of at least 500 grams.” Thus, there is ample opportunity for babies to not be counted either because they are considered nonviable or because a certificate that they were born alive and viable could not be obtained. The scale of the resulting impact on mortality rates is implied by statistics that show that under a sixth of recorded infant deaths in France take place in the initial twenty-four hours after birth, versus one-third in both the US and Canada.
Another such case is Canada. Like that US, Canada uses the WHO definition of live birth but bureaucratic complexities may mean that not every baby born alive makes it into the register of live births. One such baby is Sonja Stefnovic, born January 6, 2006 in Ontario and who, though she lived a mere 35 minutes, more than met the criteria for a live birth. Yet, when her parents asked for a birth certificate they were told none existed. The paperwork had been filed to register her death but not her birth, which, unbeknownst to her parents, had to been registered separately. As far as Ontario was concerned, Sonja had never been alive at all.
While such cases are unusual in most of Canada, they are endemic in Ontario where in the last ten years there have been over 30,000 cases of unreported births due to bureaucratic delays, confusing procedures, and a substantial fee for registering births. That adds up to thousands of missing birth records each year and constitutes 30 percent of records of babies who lived less than a year. And Ontario makes up 40 percent of births in Canada. Ontario has also consistently been late in reporting births to federal statisticians. Arne Ohlsson, director of evidence-based neonatal care and outcomes research at the University of Toronto has said that, "[b]ecause such a large proportion of the population is born in Ontario and those vital statistics are not accurate, the country's statistics are not accurate and comparisons with other countries will be inaccurate." Whether the flawed records are used in calculating national statistics or whether Ontario is omitted, the result is nonetheless a number that does not cover all the babies born in the country.
Even if you can overcome the statistical problems, there are other factors that make infant mortality rates an inaccurate tool for comparing the quality of care in different health systems. Access to and quality of care are certainly influential in determining death rates but elements related only partially, tangentially, or not at all to the medical system are critical as well. Many are cultural, like delayed childbearing, use of fertility treatment, multiple births, or a high percentage of teen births, since these trends come with higher prematurity rates. Further, mortality rates are affected by ethnic factors that differ widely –and cannot be dismissed merely as reflections of access or socio-economics.
Culture may also trickle into medical care in another realm, in influencing the willingness of doctors and hospitals to pull out all the stops to try to save even the smallest and most premature babies. In the US, the expectation is that doctors will do everything possible to keep preemies alive unless the parents object. But in the Netherlands, premature babies below 25 weeks gestation are no longer to be resuscitated because the chance of survival without serious disabilities is considered low and they are instead given palliative treatment. Other nations fall somewhere in between and in countries like the UK and Canada, the enormous cost of intensive intervention and shortages of pediatric nurses and spaces in NICUs have played a significant role in the debate over when and how to intervene.
What the ‘correct’ policy may be is beyond the scope of my present post –and the answer is probably neither singular nor fixed. But it is something that needs to be taken into account when comparing statistics.
All of these issue show why international statistics should be interpreted with extreme caution, at least until we know where the number come from, how they are derived, and what other factors are simultaneously in play. And although it may be frustrating to know that, as in this case, the answer may be that the statistics cannot tell us much about the strengths and weaknesses of health care systems and their relation to one another, it is better to know we are in the dark than to clutch at false illumination.
In other cases, bureaucracy intrudes so that even those who meet the definition of live birth may not end up in statistics. In France, in order for the baby to be given a birth certificate as a live birth, one must have “a medical certificate [that] attests that the child was born ‘alive and viable’.” On the other hand, “[i]n the absence of a medical certificate attesting that the child was born ‘alive and viable,’ the civil service officer only establishes a certificate of a child without life,” in lieu of a birth certificate. Overall, “this procedure applies, on one hand, to children born alive but not viable, and on the other, to children stillborn after a term of 22 weeks…or with a weight of at least 500 grams.” Thus, there is ample opportunity for babies to not be counted either because they are considered nonviable or because a certificate that they were born alive and viable could not be obtained. The scale of the resulting impact on mortality rates is implied by statistics that show that under a sixth of recorded infant deaths in France take place in the initial twenty-four hours after birth, versus one-third in both the US and Canada.
Another such case is Canada. Like that US, Canada uses the WHO definition of live birth but bureaucratic complexities may mean that not every baby born alive makes it into the register of live births. One such baby is Sonja Stefnovic, born January 6, 2006 in Ontario and who, though she lived a mere 35 minutes, more than met the criteria for a live birth. Yet, when her parents asked for a birth certificate they were told none existed. The paperwork had been filed to register her death but not her birth, which, unbeknownst to her parents, had to been registered separately. As far as Ontario was concerned, Sonja had never been alive at all.
While such cases are unusual in most of Canada, they are endemic in Ontario where in the last ten years there have been over 30,000 cases of unreported births due to bureaucratic delays, confusing procedures, and a substantial fee for registering births. That adds up to thousands of missing birth records each year and constitutes 30 percent of records of babies who lived less than a year. And Ontario makes up 40 percent of births in Canada. Ontario has also consistently been late in reporting births to federal statisticians. Arne Ohlsson, director of evidence-based neonatal care and outcomes research at the University of Toronto has said that, "[b]ecause such a large proportion of the population is born in Ontario and those vital statistics are not accurate, the country's statistics are not accurate and comparisons with other countries will be inaccurate." Whether the flawed records are used in calculating national statistics or whether Ontario is omitted, the result is nonetheless a number that does not cover all the babies born in the country.
Even if you can overcome the statistical problems, there are other factors that make infant mortality rates an inaccurate tool for comparing the quality of care in different health systems. Access to and quality of care are certainly influential in determining death rates but elements related only partially, tangentially, or not at all to the medical system are critical as well. Many are cultural, like delayed childbearing, use of fertility treatment, multiple births, or a high percentage of teen births, since these trends come with higher prematurity rates. Further, mortality rates are affected by ethnic factors that differ widely –and cannot be dismissed merely as reflections of access or socio-economics.
Culture may also trickle into medical care in another realm, in influencing the willingness of doctors and hospitals to pull out all the stops to try to save even the smallest and most premature babies. In the US, the expectation is that doctors will do everything possible to keep preemies alive unless the parents object. But in the Netherlands, premature babies below 25 weeks gestation are no longer to be resuscitated because the chance of survival without serious disabilities is considered low and they are instead given palliative treatment. Other nations fall somewhere in between and in countries like the UK and Canada, the enormous cost of intensive intervention and shortages of pediatric nurses and spaces in NICUs have played a significant role in the debate over when and how to intervene.
What the ‘correct’ policy may be is beyond the scope of my present post –and the answer is probably neither singular nor fixed. But it is something that needs to be taken into account when comparing statistics.
All of these issue show why international statistics should be interpreted with extreme caution, at least until we know where the number come from, how they are derived, and what other factors are simultaneously in play. And although it may be frustrating to know that, as in this case, the answer may be that the statistics cannot tell us much about the strengths and weaknesses of health care systems and their relation to one another, it is better to know we are in the dark than to clutch at false illumination.
