Main content

Health checks, Fertility, Adjustment

Dr Mark Porter asks if there is evidence that NHS health checks or 'mid-life MOTs' prevent serious disease. And are women wanting children aware that fertility declines with age?

NHS health checks or 'mid-life MOTs' have hit the headlines as new research claims they are a success. The aim is prevention - of diabetes, heart attacks and strokes - but their introduction has been controversial amid criticism they are not evidence based or cost effective. Resident sceptic Dr Margaret McCartney debates the issues with National Clinical Advisor Dr Matt Kearney.

And putting the family back into planning. As more couples leave it later before starting a family there is growing concern from fertility experts that many people don't know enough about when female fertility starts to decline. Professor Adam Balen and Professor Joyce Harper discuss the issues. And how accurate is the perception, often reported in the media, that fertility 'drops off a cliff' in the mid to late thirties? Professor Richard Anderson reviews the so called 'broken stick' study, a mathematical model which first defined the sharp drop off of female fertility.

And another instalment of Inside Language where Dr Margaret McCartney and Professor Carl Heneghan examine the terms used in evidence based medicine and why they matter. This week, adjustment and how researchers allow for factors that might skew their findings.

Available now

28 minutes

Programme Transcript - Inside Health

Downloaded from

THE ATTACHED TRANSCRIPT WAS TYPED FROM A RECORDING AND NOT COPIED FROM AN ORIGINAL SCRIPT.听 BECAUSE OF THE RISK OF MISHEARING AND THE DIFFICULTY IN SOME CASES OF IDENTIFYING INDIVIDUAL SPEAKERS, THE 成人快手 CANNOT VOUCH FOR ITS COMPLETE ACCURACY.

INSIDE HEALTH

Programme 12. 鈥 Heath Check, Fertility, Adjustment

TX:听 29.03.16听 2100-2130

PRESENTER:听 MARK PORTER

PRODUCER:听 ERIKA WRIGHT

Porter

Coming up today:听 Fertility 鈥 should we be doing more to put the family back into family planning?

And how researchers adjust for confounding factors 鈥 Carl Heneghan and Margaret McCartney are back with another Inside Language to help you interpret the real story behind the headlines.

But first NHS health checks. These mid-life MOTs have been in the news following the publication of research suggesting the seven year old programme has already prevented over 2,000 strokes and heart attacks.

The checks vary depending on where you live. In England they are offered to everyone between the ages of 40 and 74; in Scotland it鈥檚 40 鈥 64 and in Wales to anyone over 50. Whatever the age the aims are much the same - to spot problems like high blood pressure, heart disease and diabetes before they might otherwise become evident.

But their introduction has been controversial amid criticism that they are not evidence based and that they鈥檙e a costly draw on resources.听

So supporters of the programme have welcomed a new study assessing the impact of the checks in England which seems to suggest that they do work 鈥 albeit at a price.

Dr Matt Kearney is a GP in Runcorn and National Clinical Advisor to both NHS England and Public Health England.

Kearney

First of all the people come in for the health check were from all sectors of the community, so people from deprived communities were more likely to attend than people from affluent communities.听 That鈥檚 very reassuring, it鈥檚 early evidence but the early evidence suggests that we鈥檙e not just attracting the worried well, which I think was an important potential criticism.

Secondly it showed that individuals coming into the programme are having risk factors identified.听 I鈥檓 thinking about obesity, smoking, high alcohol intake.听 Significantly more numbers are getting detected and people are being referred on and offered evidence based interventions.听 But I think very importantly what it鈥檚 also picking up is people who鈥檝e got undiagnosed serious physical risk factors, so by that I mean things like high blood pressure, diabetes, atrial fibrillation 鈥 conditions that significantly increase the risk of having a heart attack or stroke that we鈥檙e not picking up in general practice, despite having one of the best primary care systems in the world and clearly it is, and there are gaps, we鈥檙e managing to miss significant numbers with those conditions.听 So the evidence shows us that from this early evaluation one in 27 health checks is identifying someone with previously undiagnosed hypertension, about one in 110 with previously undiagnosed diabetes.听 Those are very important findings for those individuals who then go on to appropriate management of those conditions which might include drug therapy but treatment is going to significantly reduce their risks of having heart attacks or strokes.

Porter

Dr Margaret McCartney鈥檚 in our Glasgow studio.听 Margaret, you鈥檝e been looking at that evaluation, what was your take on the findings?

McCartney

Well the evaluation is all very well but it鈥檚 an uncontrolled reporting of what happened and it鈥檚 not able to say what the difference health check made because it doesn鈥檛 compare it, again what would have happened if we hadn鈥檛 done anything like that at all because it wasn鈥檛 a randomised control trial.听 And we do know that what happens when you go and look in a community for a disease or risk factors you will find it but the question is whether you find it usefully earlier than what would have happened had you just left things to its own devices.听 So this study isn鈥檛 good enough to tell us whether or not health checks have done what Jeremy Hunt Tweeted to say was, in his words, 鈥済reat news that NHS health checks have now saved hundreds of lives, 2,500 heart attacks and strokes avoided thanks to early intervention鈥 鈥 we can鈥檛 say that on the basis of this study.听 And in addition we would have to assume that of the many people who attended, even though that was only 12% overall, you would have to assume that the vast majority of them went on and took their medication as prescribed for a long time and it had the effect that鈥檚 been shown in trials already.听 So these are extrapolations of data rather than real life data where we know it鈥檚 made a difference.

Port

Matt, your take on that?

Kearney

Yeah, this is not defining evidence, this isn鈥檛 a randomised control trial that proved whether or not this approach to an NHS health check is going to deliver what we hope it will deliver, that鈥檚 reduction in premature death for the population.听 None of the evidence that鈥檚 been published to date answers that question for us and much of the studies 鈥 the Cochrane Study that鈥檚 been discussed for example 鈥 that some people flag up as proving that health checks don鈥檛 work actually most of those studies were from the 1960s, all but one before 1992, and before the era of modern medicine and the modern approach.听 So I would agree we don鈥檛 actually have that hard evidence that shows this programme is going to work.听 What we do have is evidence that the interventions contained in the programme, they鈥檙e all absolutely evidence based.听 But I frankly would take it beyond that 鈥 I acknowledge that we have a gap in the evidence and I think we need to plug it.听 I guess where I鈥檓 coming from Margaret, as a GP like yourself, is that we鈥檙e not in a position where we can wait for perfect evidence because another stream of evidence that we have is the evidence that we have a relentless rise in preventable mortality because of lifestyle issues, because of undiagnosed hypertension, undiagnosed diabetes 鈥 that鈥檚 something that鈥檚 going to break the NHS bank, we don鈥檛 have the luxury of waiting for perfect evidence.

McCartney

But the problem is if we don鈥檛 try and get better evidence we simply just throw good money after bad.听 And there have been two big studies that are of real relevance in this situation.听 One was the Inter99 Study which was a very similar study to the health checks programme, which was a randomised control trial, it was done on the continent and it looked at a community based individually tailored programme for screening for the risk of heart disease and for risk factors for premature death in the community.听 And if found that it did not work over 10 years, that was a really high quality trial.听 The other trial that I think is of relevance as well was a UK based trial called the ADDITION-Cambridge Trial, people who were at high risk of diabetes were screened for it and they were offered interventions earlier for it and that did not reduce down the mortality rate to do with any all-cause mortality, cardiovascular mortality or diabetes related mortality.听 And the problem is that when we keep doing the same thing again and again and expect different results we鈥檙e going to fail, we have to acknowledge the fact that we actually do have really good recent modern evidence that says that this approach doesn鈥檛 work and stop treating lifestyle factors as something that needs an individual doctor or nurse to intervene on and start looking at the public health interventions that we know that work.

Kearney

Well I do acknowledge that the Inter99 Trial is clearly an important trial but there are major differences between the studies they looked at and the NHS health check and it鈥檚 not strictly comparable 鈥 younger population, lower risk factors etc.听 But just to pick up the other points that Margaret has made.听 Again I agree Margaret, I think population level measures, measures by government, local government, employers, schools 鈥 these are all the big drivers that are going to help reduce preventable illness and they have a very important part to play.听 But what I would argue also is that we have an important part to play as GPs, nurses, people working in the NHS and I think this helps us because it gives us a structure to our prevention conversations with patients and it is about those conversations.

Porter

Margaret, how do you respond to that fact then that this is just facilitating a discussion that perhaps should be taking place anyway?

McCartney

But we do do it anyway and this is my point.听 These conversations go on up and down the country, every single day GPs are talking to patients about what would you like, what would you not like, what are the pros, what are the cons 鈥 that is what we do.听 The point is whether or not formalising this in the system of health checks, which gets loads of money thrown at it, creates waste, creates duplication, whether that itself is a useful waste of money and I argue that it鈥檚 not because we do not have the evidence behind it.听 Health checks, I think, offer promises that I don鈥檛 think are actually realistic at the end of the day.

Kearney

Well I think that鈥檚 where we differ.听 I mean I think it is a very reasonable expenditure by the NHS showing that it鈥檚 getting serious about prevention, it鈥檚 allied to lots of other things that we鈥檙e doing in the NHS like the diabetes prevention programme, the idea of a sugar tax across the NHS, the healthy towns 鈥 these are all big things, it鈥檚 part of a bigger picture.听 And I agree we don鈥檛 have the evidence to answer the fundamental question 鈥 is this going to work in the long term 鈥 but we don鈥檛 have the luxury of waiting.

McCartney

We can鈥檛 ignore the evidence that鈥檚 already there that just says this is not an evidence based use of money鈥

Kearney

The evidence at the moment is incomplete or irrelevant and that鈥檚 the issue鈥

McCartney

No I don鈥檛 think you can call Inter99 or ADDITION irrelevant鈥

Kearney

I do agree that we need鈥.

McCartney

What鈥檚 the point in doing these big enormous studies when we just ignore them when we don鈥檛 like the answer.

Kearney

I think it鈥檚 about taking a balanced approach to the evidence that鈥檚 going to help us make a decision鈥

McCartney

鈥 the problem is this is a fig leave, having the health checks sort of set in statute essentially in England, which is what it is, it creates a fig leave for the government to say oh we鈥檙e doing something about this and I don鈥檛 think that鈥檚 good enough because I think you have to say this doesn鈥檛 work and therefore you absolutely must do the big population level issues that do work, that to me is the problem when we keep pretending that this is a solution when we鈥檙e actually just delaying it.

Porter

The other concern I鈥檝e had raised with me is that it makes people feel like a patient, you鈥檙e turning healthy people into patients often on some form of medication as well.

Kearney

Sure I think that鈥檚 a very important question.听 I think we do need to be very cautious about potential harms that we do for our patients and that鈥檚 our role as GPs, we鈥檙e making those sort of judgements all the time and supporting people to choose between two options that are maybe not that palatable and it鈥檚 very important to reduce harm.听 And it is about those conversations, it鈥檚 not about putting people on medication primarily.

Porter

You describe the evidence at the moment as not being at the level that you require, you鈥檙e taking a pragmatic approach you鈥檙e saying but if it turns out in a few years that this is not cost effective or is not effective will it be stopped?

Kearney

Absolutely, we鈥檙e going to take a hard look at the evidence and yes.

Porter

And how long will that be do you think before we have the sort of evidence that would satisfy someone like Margaret for instance?

Kearney

Well I don鈥檛 have the answer to that question.听 NHS England and Public Health England established an expert scientific clinical advisory panel and that committee will take expert views on the available evidence and as evidence emerges yes we鈥檒l give an honest appraisal and an honest review as to what the evidence shows.

Porter

Dr Matt Kearney.听 And there is a link to the evaluation of the first four years of the programme in England on the Inside Health page of the Radio 4 website.

Now, how much do you know about fertility? 听As more women leave it later before starting a family there is growing concern that many don鈥檛 know enough about when female fertility starts to decline. And by the time some do find out, it can be too late. While we put a lot of effort into CONTRAception, we often take CONception for granted.

Adam Balen is Professor of Reproductive Medicine and Surgery at the University of Leeds and one of a number of leading fertility experts keen to put the family back into family planning.

Balen

Well I see this every day in my clinic that couples come and see me in the infertility clinic, having stopped contraception thinking that they鈥檒l get pregnant straightaway and I don鈥檛 think we鈥檙e saying enough about the impact that getting older has on fertility.听 So fertility declines significantly as you get older because women are born with a fixed number of eggs and they are lost at a relentless rate, even if you鈥檙e taking the contraceptive pill or any other form of contraceptive you鈥檙e losing your eggs and therefore it鈥檚 essential that we鈥檙e informing our young couples that they shouldn鈥檛 leave things too late, that planning for a family is not only about contraception and safe sex when you鈥檙e very young but also thinking a little bit further ahead 鈥 when are you going to want to start a family, don鈥檛 leave it too late.听 I鈥檓 talking here really about education and actually bringing that education right back to teenagers and our youngsters who need to be becoming aware, even though it may not be at the forefront of their minds, they need to be told that this is an issue.

Vox pops 鈥 Two teenagers

I鈥檓 13.

And I鈥檓 13 too.

I want to start having kids when I鈥檓 about 26, 27 because if there鈥檚 something wrong with my body that can be helped I want to not leave it too late.

I just want to have like a secure job and everything first and then think about it, but don鈥檛 leave it like after 30.

I want to have a stable career, earning a reasonable salary and definitely be in a stable relationship as well.听 So I鈥檇 want to start when I鈥檓 like 27 and have my first kid, if all good, about 28.

I don鈥檛 want them more than two or three years apart, that they can鈥檛 like play with each other.

Yeah exactly.

Balen

We need to support our youngsters to achieve all that they desire.

Porter

Adam, looking at the couples you see in your clinic, I mean are there really a significant proportion of them who didn鈥檛 know that delay might cost them their ability to start a family?

Balen

Yes sadly people are surprised because these days everybody expects to have control over their lives.听 We鈥檙e very good at controlling everything but we can鈥檛 control our biological clocks.听 When you stop the pill you think you鈥檙e going to get pregnant next month but in reality it can be many months and sadly for some it just doesn鈥檛 happen.

Harper

My name is Professor Joyce Harper and I鈥檓 a professor at University College London and I work in the Department of Reproductive Health.听 As we get to 30, 35, 40 it gets harder and harder biologically to have a child, our biological clock is really ticking but women do not understand that.听 They come to our IVF units aged 38, 40, 45 and think that we can make them a baby and we can鈥檛.

Porter

But whose fault is that, it鈥檚 not a new message, I mean it鈥檚 been around for decades, why is it a surprise to so many women?

Harper

The message hasn鈥檛 got out there unfortunately and we really need to do something about it and it really needs to be set in our schools.听 Some of the major education documents that have come out at the moment they don鈥檛 even mention fertility, they mention contraception, they mention sexually transmitted diseases but not fertility.

Vox pops 鈥 Two teenagers

I think fertility鈥檚 something to do with children and like babies.

I think it鈥檚 to do with if your eggs like鈥

If you can have kids.

Yeah, yeah I think that鈥檚 what it means.听 Because the word fertile kind of sounds like鈥

Fertility rate 鈥 I鈥檝e heard something like how many or if you can have kids as well, I think it鈥檚 something to do with that.

Yeah something, we don鈥檛 really talk about it.

No, you learn about it like in science and stuff but they only like say oh fertility rate and they go on to something else, they don鈥檛 really explain it.

They don鈥檛 explain it, people don鈥檛 think it鈥檚 very important to know about it.

Harper

We need to apply this to boys and girls, we need to make boys become fertility aware as well and make them realise that they can鈥檛 wait until they鈥檙e 50 if their partner鈥檚 the same age as them to have their first child.听

Porter

You hear this subjectively from a lot of people working in your field but is there any hard evidence that women really don鈥檛 understand as much as you鈥檙e suggesting?

Harper

No you鈥檙e right, we haven鈥檛 got hard evidence, we鈥檝e got our anecdotal evidence from our experience as IVF practitioners when we get patients coming in and from friends and family.听 But one of the real problems is that we know that women are delaying their first conception, that data for many countries is out there and we can see year in year out that it鈥檚 increasing.

Woman

I鈥檓 29, I work as a government economist and I鈥檓 in a relationship, I鈥檝e been with my boyfriend for about a year and a half.听 Yeah I would like kids, I鈥檇 like two probably in the next five years 鈥 we鈥檇 like to go travelling first, it would be great if I got promoted.听 I鈥檇 like to get married.听 I鈥檇 like to buy a house at least, I mean my boyfriend and I don鈥檛 even live together, hopefully before 35 but I can just see the time ticking away, I can completely understand how that happens for lots of people.

Harper

I think nowadays women are doing so many other things, we鈥檝e got so many other options compared to 20 or 30 years ago, so even if they are aware of their biological clock I think it鈥檚 at the back of their mind and they鈥檙e not thinking about it, then when they do settle down they realise that actually we may have a problem now.

Woman

But I have a lot of girlfriends who are the same age as me 鈥 29 鈥 only out of I鈥檇 say 30 one鈥檚 had a baby and one鈥檚 thinking about it.

Porter

And the rest of that peer group 鈥 assuming they want children 鈥 are taking a risk. Or at least that is what coverage in the media would have you believe as it often suggests fertility drops off a cliff in a woman鈥檚 mid to late thirties. But is the decline really that sudden? Richard Anderson is Professor of Clinical Reproductive Science at the University of Edinburgh

Anderson

One of the holy grails in reproductive medicine historically and in fact it still is is knowing really how many eggs a woman has in her ovaries.听 You ovulate one egg every month and it鈥檚 been known for a long time that the number of follicles, which are the structures containing the eggs, is actually established before birth in women and then progressively declines.听 And the problem is that really the only way to actually assess that accurately is to take an ovary out of a woman and cut it into thin slices and count the follicles under the microscope because they鈥檙e too small to see by any scan or other imaging and they are no real biochemical tests that really give you such an accurate picture.听 So what happened was that a number of samples were collected over the years, there was a German who did a lot of this work in the 1950s and then a Frenchman called Alain Gougeon who did a fantastic amount of work in the late 鈥70s, early 鈥80s.听 And then it was really putting all these data sets together by a mathematician called Malcolm Faddy in combination with a biologist called Roger Gosden which 鈥 there was a beautiful meeting of minds and skills 鈥 who started analysing this in a more mathematical way to put together the picture that we are so familiar with today of this increasingly steep decline in the number of eggs that a woman has in her ovaries over the years.

Porter

So this was a mathematical model based on real data coming from samples taken from women?

Anderson

That鈥檚 right, it was looking down the microscope, counting the eggs in ovaries that had been removed from women across life for a variety of different reasons, as you can imagine.听 And then Malcolm and Roger demonstrated that there was an increasingly sharp falloff in the number of follicles, the number of eggs within the ovary, with age and their mathematical modelling described this best as two straight lines which was effectively a broken stick 鈥 so there鈥檚 one rate of decline across a certain range of ages up to the late 30s and thereafter it gets steeper.听 And where those two 鈥 drawing those lines on a graph effectively looks like a broken stick and that鈥檚 been the shorthand for it ever since.

Porter

And that acceleration in the decline in fertility occurred at what age, according to them?

Anderson

Around the age of 37, 38 was when the cut off seemed to happen.

Porter

And has that model proved accurate?

Anderson

It鈥檚 proved fairly accurate, although I think people have realised now that actually there鈥檚 no such sharp cut off and it鈥檚 more of a continuous process.听 So what鈥檚 happened over the last subsequent 鈥 actually probably about 30 years now, so this has been over a very long time, is that more data sets have been added to the data that Malcolm and Roger originally had, so you can get a more accurate mathematical analysis.听 And actually the most recent one published was done by another mathematician colleague of mine here in Scotland, from St Andrews, which shows that really it鈥檚 more of a steepening curve rather than actually a broken stick.听 There is no such thing as suddenly falling off a cliff, this is all a process that happens over time in every woman.听 So it does get sharper with age, the time at which it starts to get sharper and steeper will vary from one woman to another, depending on her underlying biology of her ovary.听 But it doesn鈥檛 suddenly go away overnight, so these things are always a gradual process.

Porter

Because you could be forgiven for thinking, after reading a lot of the coverage that this area gets, that you鈥檙e okay until you鈥檙e 35 and then it鈥檚 all over.

Anderson

Yeah absolutely and I think there鈥檚 a lot of misinformation out there on this.听 Certainly I think the message that women need to take home is that their fertility does decline with age and modern medicine has very limited ability to compensate for the effects of time.听 But there鈥檚 certainly no absolute age at which suddenly things switch off at all for any individual woman.

Harper

There鈥檚 a lot of data showing that if you want three children you really need to start in your late 20s.听 So unfortunately evolution hasn鈥檛 caught up with social lives, so socially we鈥檙e delaying childbirth in every country but our menstrual cycle is still meaning that we鈥檙e going to go through the menopause at the same age we would 50 years ago, nothing鈥檚 changed biologically.

Porter

What do you think of the practice of women assessing their ovarian reserves 鈥 so they go along, they have a test 鈥 a scan or a blood test or both 鈥 which tells them what state their eggs are in?

Harper

There has been some publicity about this and in my view the test that we have at the moment 鈥 it tells us some basic information and it certainly helps us when they鈥檙e going through fertility treatment to decide what treatment they have but really being a solid indicator of when their fertility is going to decline it鈥檚 not there yet.听 So there are other things they can do 鈥 they can, for example if they have a partner 鈥 check their sperm count 鈥 and they can check their hormones and some other things to see if there鈥檚 any major problem but at the moment there鈥檚 no test that鈥檚 really going to say okay aged 28 I do this test I鈥檓 going to be okay till I鈥檓 40 or I鈥檓 not going to be okay, better start at 35.听 We haven鈥檛 got there yet.

Anderson

I think the important thing is that there鈥檚 no such thing as a fertility test, you can鈥檛 come along and say yes you鈥檙e fertile.听 But increasingly what people have recognised is that there are tests, they can give you some idea of how many eggs you have left in your ovary and there鈥檚 a blood test that can contribute to that and also an ultrasound scan looking at the ovary in more detail and counting the follicles can give you an index of that.听 But they鈥檙e really very poorly related to what your chances are of actually conceiving either naturally over the next six months, year or so, or actually through IVF.听 So there are a couple of studies that in younger women show very clearly that what these tests show has no prediction of the chances of getting pregnant over the next six months.听 And even in IVF where they do predict the number of eggs you鈥檙e going to get in an IVF cycle the relationship with actually whether it鈥檚 going to turn into a live baby to take home at the end of the day is actually fairly weak.

Porter

So a good ovarian reserve may be falsely reassuring?

Anderson

It鈥檚 clearly a good thing but it鈥檚 probably more related to the quantity rather than the quality and it certainly isn鈥檛 something to rely on because none of the tests that we have can tell you anything about the quality of the eggs that you have.

Porter

Imagine a couple are unable to conceive naturally and they go for assisted techniques, so they鈥檙e having IVF, what鈥檚 the impact of maternal age on that?

Anderson

There鈥檚 still a very significant effect of maternal age on IVF and if you look at the HFEA database which is publicly available for all clinics you can see how much that changes with age with really very poor success rates into the 40s compared to the early 30s.听 So it鈥檚 not something that modern medicine can really cure by a long stretch of the imagination.听 And a lot of the issue that we see these days in the infertility clinic is we don鈥檛 find anything particularly wrong with a couple, just as we didn鈥檛 10 years ago, but for many cases it鈥檚 just really the woman鈥檚 age is becoming the dominant factor.

Porter

Professor Richard Anderson. More details on our website, where you will also find a link to the Inside Health podcast to have it sent to you automatically so you need never miss another episode again.

Now time for some Inside Language 鈥 the series that demystifies the terms used by researchers and those reporting their conclusions. This week Dr Margaret McCartney and Carl Heneghan, Professor of Evidence Based Medicine at the University of Oxford, discuss adjustment and how researchers allow for factors that might skew their findings.

Heneghan

This is quite a simple idea but can confuse you in some ways.听 So I鈥檓 going to give you an example where you roll a ball on two different surfaces.听 You roll a ball on a carpet and then you decide to roll it on the vinyl floor in your kitchen.听 And you would expect to see that the ball would travel further on the vinyl floor but when you measure it you get the opposite answer and it goes further on the carpet and you鈥檙e mystified.听 And you go back and you look at it and you see there鈥檚 a different variable that you鈥檝e not accounted for 鈥 that the floor in your carpet in your living room, for instance, has actually got a slope on it and it鈥檚 sloping down and that slope has accounted for the ball travelling further.听 And say it鈥檚 five degrees.听 Well when you adjust you have to take in account that slope and the mathematical techniques for doing that are trying to get rid of the five degree angle to bring it up to zero, so you get a true picture of how far the ball goes.听 And that鈥檚 what you do when you call adjustment.听 But in many life situations and health situations there are more one variable, so you have to take into account two, three and four and then it gets more complex and often you can鈥檛 take account of all of them.

Porter

I suppose, Margaret, the simplest thing that we might do is to adjust for age between two groups that you鈥檙e comparing because so much in health is related to age.

McCartney

Yeah and sometimes sex as well, so sometimes studies are adjusted in account of having more men in one group compared with females in the other.听 So yes there are and I suppose I am always a little bit suspicious of adjustment because it鈥檚 another area of potential error really.听 I鈥檓 always a bit of a worrier when I see that lots of adjustment has been occurring within a particular study because I wonder how good was the study set up to be done in the first place that you had to make so many more different adjustments.

Porter

How robust is the science behind the more common adjustments that are made Carl, do you trust them?

Heneghan

It鈥檚 easy when you鈥檝e got one simple variable, like the slope, and then it鈥檚 simple to explain.听 But once you add in more and more variables it starts to break down.听 And in some areas of health, in some areas of risk, we are inherently complex in what we do on a daily basis, our interaction with the environment and trying to disentangle all that is incredibly difficult.

McCartney

I always feel as though statistical adjustment is a bit of a dark art and when it comes to this I really sometimes feel as though it鈥檚 a bit of voodoo medicine, how much can you really reliably start to adjust your data.听 And I think sometimes lots of adjustments happen because you didn鈥檛 have good randomisation to start off with and can you be sure that your statistical adjustments can make up for that fact.听 And I鈥檝e been reading lots of statistical texts, so obviously I鈥檓 not a statistician, who are fairly clear in saying that observational studies are not a substitute for randomised control trials no matter how good your statistical adjustments may seem.听 And I think this is a problem 鈥 a lot of the times we鈥檙e using statistical adjustment to make up for the fact that we鈥檙e doing observational non-randomised trials which are inherently more unreliable and we鈥檙e trying to make them more reliable using these mathematical techniques.听 And I鈥檓 sure they are valid in many situations but I worry that sometimes we end up with what seems to be quite firm conclusions where they鈥檙e actually resting on quite a lot of sand when actually we should be questioning it a bit more.

Porter

Carl, do you think our listeners should be suspicious if they see lots of adjustment?

Heneghan

I鈥檓 totally with Margaret here but there are some times when we see no adjustment that I get really worried.听 So it鈥檚 a bit of a double edged sword if you like.听 I鈥檒l give you an example.听 If I just said to you well let鈥檚 look at the relationship of blood pressure to stroke and we did that and we didn鈥檛 account for age, as you said before, we could come up with a very odd answer if we just looked at 20 year olds or 30 year olds.听 So there鈥檚 an important aspect that first we have a spectrum of patients, that you will see the disease likely in and two is you do have to account for some factors because we know as you get much older the risk of stroke gets much higher.听 And I think that鈥檚 important.听 But the second aspect is when you get to a certain point, at some point you have to go okay I鈥檓 going to call it quits here.听 And what鈥檚 happening in the modern world is people have much more access to data now, data mining is the word, where you can just go looking for these associations and make it plausible by saying I made one or two adjustments.听 And often that is sloppy science.

Porter

Professor Carl Heneghan and Dr Margaret McCartney. And that鈥檚 it for this series of Inside Health. We will be back in July. Until then, goodbye.

END

Broadcasts

  • Tue 29 Mar 2016 21:00
  • Wed 30 Mar 2016 15:30

Discover more health facts with The Open University

Can you detect health fact from fiction?

Podcast