So much of life is predicated on a capacity to predict. We want to know what the weather is predicted to be tomorrow, this week, next week. We want our doctor to predict the likely progression of our illnesses and the effect of treatment. Investors would like to know what the stock market is likely to do this week, this year. All of us would like to know if climate change is a reality, and if so, how it was generated and how fast it is predicted to progress. And at this time of financial uncertainty governments need to predict where the economy is heading, and assess the likely outcome of its legislative moves. Accurate prediction is crucial. [more]
To predict with optimal precision, expert opinion is needed. Although predictions in the field of politics affect profoundly the way politicians think and vote, the quality of these predictions has been disappointingly variable – from the very poor to the best available. As high quality research is not always available, decisions are sometimes made on what is available. When that is so, there is a need to be aware of its quality. It is a matter of concern that sometimes those who use research of indifferent quality for decision-making are either unaware of its quality or indifferent to it, so long as it serves their purpose. In political circles there is a regrettable propensity to embrace research that endorses a favoured position, irrespective of its quality.
In his Monday piece on Pollytics titled Nostradoofus, Possum addresses this issue by way of reference to two Government moves, changes to the threshold for the Medicare levy, and the ‘alcopops’ tax.
He draws attention to the hullabaloo that engulfed the change to the Medicare levy threshold whereby the salary at which the levy became operative for people without private health insurance rose from $50,000 to $70,000 for singles, and $100,000 to $140,000 for couples. There were dire predictions from the health insurance industry, the AMA and the Coalition of a mass exodus from private health insurance, with resultant severe overcrowding of public hospitals, particularly their emergency departments. These predictions did influence politicians and their voting behaviour.
Yet a piece in The Australian by Siobhain Ryan on February 18 titled Private health funds defy exodus fears showed that “...private health funds had recorded their highest membership rates in seven years, defying warnings that controversial changes to the Medicare levy surcharge would cripple the industry. The percentage of Australians holding hospital cover with private health funds rose to 44.8 per cent at the end of December, up from a revised figure of 44.7 per cent three months earlier.” So much for the dire adverse predictions. On what were they based?
Yesterday Nicola Roxon announced that health insurance premiums would rise by 6 per cent on April 1, an amount well below Joe Hockey’s prediction of a rise of 12 per cent which he insisted would result from the Government's changes to the Medicare levy. On what evidence was that prediction based? A premium rise is not surprising; it is needed each year to cover increased health coverage, increased benefits paid to patients, rising health service costs, and this year investment losses from the global financial crisis. Despite the increase being lower than during the last five years under the previous government Peter Dutton decried this new Government ‘tax slug’ on families, which of course it is not.
So here is just one example of predictions being faulty on several fronts, probably because they were based on inadequate research. Yet politicians’ judgements, decisions and utterances were influenced by that research. One wonders how much scrutiny was applied to the research that so confidently predicted such catastrophic sequelae, predictions that were widely promulgated by those opposed to the changes to the levy.
The ‘alcopops’ tax is another example. It was opposed on the grounds that it would not ameliorate binge drinking among young people, particularly the young women who were consuming so much of these drinks. The research that was used to support that conclusion was not clearly articulated, it was not subject to rigorous scrutiny, it was just accepted as valid even although parallel situations were few and far between. Remember Brendon Nelson, who previously supported doing something about binge drinking, predicting the dire affect this legislation would have on his ‘ute drivers’ who apparently habitually consume Bundy and Coke. He should have known better than accept paltry evidence. The Coalition labelled it as just a ‘tax grab’ with no prospect of it resulting in any health benefit.
Then came the time for review of the alcopops tax, and its confirmation by legislation. So was it effective? The very day it was due for consideration, Access Economics came out with a study commissioned by the people who make alcopops that showed that the tax had made no difference to binge drinking. But when questioned how that conclusion was reached it emerged that it rested on a single diagnostic code applied in public hospital emergency departments. Quickly medical researchers pointed out that using that criterion alone was a flawed approach. Nicola Roxon was more explicit. She said the study used data over such a short time period that the researchers admitted that made it unsuitable to base firm conclusions on, that the hospital data was of dubious quality, that as the report itself pointed out: "there is no agreed national approach to collection of diagnosis codes and demographic information for ED patients", and that it used just one diagnostic code for most of its data, excluding others and distorting the statistics. Moreover she said it did not show any causation between the Government's actions and the effects it alleged to show, and ignored ATO figures on alcohol consumption. So much for the Access report, which on the face of it appears to be shonky, compromised not just by methodological flaws but by being sponsored by the distillers who make alcopops. A study by Professor Tanya Chikritzhs, of the National Drug Research Institute at Curtin University that compared industry-provided data for the three months to June 2007 with the same period in 2008 found that “Australians consumed 91 million fewer standard drinks in the Ready-To-Drink (RTD) category in the three months following the tax hike - a drop of 26 per cent on the 2007 period.” She said “The argument that has been made so far is this (tax hike) won't work because young people will move on to something cheaper,...There does appear have been a shift to beer and spirits, but when you total up that increase it only accounts for about half the decline in RTDs”.
So there you have it, a distiller-funded Access Economics study that showed the alcopops tax ineffective, and a National Drug Research Institute study by an academic in the field of drug use that showed the tax reduced alcopops sales by a quarter and only half of that was compensated by substitution – in other words alcohol consumption fell.
These two case histories in the health field illustrate how important it is to base decisions on hard verifiable facts unearthed by sound research, and the danger of poor quality studies that do not meet the criteria for reliable research. Good social research is difficult, but not impossible. In science and medicine there is a discipline called ‘critical appraisal of research studies’ that involves, among other things, appraisal of methodology, sampling, data analysis and the conclusions that have been drawn from the findings. There is a pressing need to apply critical appraisal to any research that lobby groups present to support their case, and to the research politicians use to form their views, to make decisions, and to support their voting.
If this country is to be governed by parliamentarians who base predictions, decisions and actions on valid and reliable evidence, a much more rigorous appraisal of that evidence is now needed. To do otherwise is dangerous and dishonest.