The Best Australian Science Writing 2015 (27 page)

I, wormbot: the next step in artificial intelligence

Uncharted waters

An uneasy alliance

Honest placebos

Jane McCredie

There is one treatment for conditions ranging from depression to back pain to irritable bowel syndrome that produces impressive results with minimal side effects. Yet clinicians are often reluctant to prescribe it.

Why? Because this particular silver bullet is the placebo effect.

Although the mechanisms of placebo remain largely mysterious, the belief that we are receiving a helpful treatment somehow seems to have the ability to make us feel better.

And there's the rub. If we need to
believe
in it for placebo treatment to work, then surely the clinicians dispensing it would have to practise a certain level of deception.

But maybe it's not that clear cut.

Canadian psychologist Dr Susan Huculak argues we need more open discussion about the role of placebo in medicine as well as clear guidelines for clinicians on how and when to use it.

Competing viewpoints on placebo – from scorn to veneration – have impeded progress, Huculak writes in an opinion article that uses psychiatry as a case study.

Clinical trial researchers tend to see placebo as no more than a contaminant, and they strive to design trials to minimise its
effects, she writes, quoting one research team writing that the placebo ‘problem' is ‘probably the most common reason for negative trials'.

So it's not the ‘drugs not working as they're supposed to' problem that's at fault here then. Phew.

In any case, Dr Huculak asks, should researchers be seeking to reduce the placebo effect in the first place?

‘If the effects following placebo are as real and potent as the drug effects (i.e. are legitimate), it would make little sense to try and reduce them', she writes.

* * * * *

This may be particularly relevant in psychiatry where a number of studies have shown antidepressants are no more effective than placebo in mild to moderate depression, despite showing efficacy in severe depression.

Given the known side effects of antidepressants, should we be exploring placebo as a treatment in less-severe forms of the disease?

Clinicians, Dr Huculak writes, are ‘generally uncomfortable or even baffled' when it comes to the placebo effect.

Despite that discomfort, several studies have suggested many doctors do use placebos in clinical practice – whether it's a recommendation of vitamins, an unwarranted prescription for antibiotics, or a subtherapeutic dose of antidepressants.

One of the reasons for the enduring popularity of alternative health therapies may be that the practitioners are particularly skilled at stimulating a placebo response.

Deliberate use of a placebo without the patient's knowledge raises clear ethical issues, but is deception always necessary for the placebo effect to work?

It seems extraordinary, but maybe not.

Researchers from the Harvard Medical School conducted a controlled trial in irritable bowel syndrome, with placebo as the active treatment. The control in this case was no treatment at all, though participants in both arms received the same level of interaction with providers.

Participants in the active treatment (i.e. placebo) arm of the trial were told they were being given ‘placebo pills made of an inert substance, like sugar pills, that have been shown in clinical studies to produce significant improvement in irritable bowel syndrome symptoms through mind–body self-healing processes'.

Those given the sugar pills experienced significantly greater improvements in their condition than those who did not get the placebo.

So there's the question: could openly giving a placebo, accompanied by the information that placebos have been shown to work, offer an ethical way to harness this powerful tool?

Will a statin a day really keep the doctor away?

Why aren't we dead yet?

Revisiting Milgram's shocking obedience experiments

Imagine there's new metrics
(it's easy if you try)

Jenny Martin

Academia has become obsessed with metrics. Institutions jostle for the ‘top' positions in international rankings, departments are evaluated nationally to identify the ‘best', and individuals are lined up against one another to find the ‘leaders'.

Let's take international rankings as an example. The rationale behind rankings such as the Times Higher Education World Rankings, the QS World University Rankings and the Shanghai Jiao Tong University Academic Ranking of World Universities is to allow people to make informed decisions about where to study, teach and conduct research. It follows then that a higher rank will mean more students, especially international students, which in turn means more money coming into the university.

Indicators used to calculate these rankings include things like academic reputation, research income, number of publications, number of citations to these papers, industry income, and the ratios of faculty:student, international:local student, international:local staff, and doctoral:bachelors student. Data is gathered from detailed surveys sent to academics, employers, and universities, as well as from companies that specialise in
providing research citation data (Thomson Reuters, Scopus etc.).

There are two things that are troubling about these indicators. The first is that there are no upper limits where there probably should be. Considerable effort is expended by institutions to increase each indicator with the aim of getting a top spot. If we extrapolate, without applying upper bounds, what could be the consequences of behaviours driven by these metrics?

Hmm, let's see. If the number of PhD students is a key indicator and there is no upper bound, this might lead to oversupply. In turn, this might result in PhD student disenchantment with academia during the course of their studies. If large numbers of PhD graduates are being produced, it's likely that a considerable proportion of early career researchers will be unable to find positions in academia. The pressures on PhD students to focus on producing research papers would lead to a lack of time and opportunity to ‘explore (other) aspects relevant to their future career options'. This situation might produce a generation of highly intelligent, highly qualified PhD students/early career researchers who feel like failures.

The second thing that troubles me is that international rankings are meant to identify the best workplaces, yet none of the rankings evaluate important indicators like job satisfaction, work–life balance or equal opportunity. Taken to the extreme, the research quality indicators might drive behaviour that leads academics to work ridiculous hours, taking no time off during the year, not celebrating their successes, and expecting the same work ethic from their team. In this scenario, leaders of the largest research teams would thrive, because they would produce more results than smaller groups, so that we might see the rise of academic Ponzi schemes.

With funding cuts to research and a growing number of large teams led by senior researchers, we might also see grant funding success fall to record lows and junior researchers miss
out on funding. Those who work best in collaborative, cooperative settings will become disenfranchised and demoralised in the hypercompetitive environment that develops. Researchers might consider cutting corners, and the academic pipeline will likely leak first with those who have significant domestic and caring responsibilities. Perhaps we might observe an increase in mental health issues among academics.

Does this set of circumstances seem familiar to anyone else?

While there are good reasons to evaluate research quality and impact, it is inevitable that bad things will happen when no checks are placed on how the loftiest research heights are attained. If the goal of international rankings is truly to identify the best places to study and work, then new metrics are needed to identify institutions that combine achieving research and teaching greatness with offering the ‘best' diversity in their professoriate, boasting the ‘top' work–life balance, and supporting ‘leaders' who train research students for positions outside academia.

With these thoughts in mind, I've dreamed up a few new metrics to use alongside the more traditional ones. Maybe this combination might lead to rankings that identify the most successful, most highly productive higher education training grounds and workplaces that are also best at supporting career aspirations and mental, emotional and physical wellbeing.

1. The no-asshole rule

At the Science in Australia Gender Equity Forum in Canberra in late 2014, I heard about the no-asshole rule. Assholes: we've all met them. We've all had to work with them. According to Bob Sutton who wrote the book on the no-asshole rule, assholes are defined by two characteristics:

•  after encountering the person, people feel oppressed, humiliated or otherwise worse about themselves

•  the person targets less powerful people.

If these characteristics apply, there is your asshole (figuratively speaking). Bob even outlined an asshole scoring system (ranging from levels 0–3) and management metric.

It is my view that in order to be eligible to participate in the new international rankings, universities must teach the no-asshole rule to all first-year students, employ no level 3 assholes and tolerate no level 3 asshole behaviour by staff or students (for one of too many recent examples, see the Dalhousie ‘gentleman' student dentists). One asshole incident without consequences means the whole institution gets the big red dislike button.

2. H-index

No, not that one. This is the Happiness index. Bhutan has one; it's called ‘gross national happiness'. I propose that staff and students at universities should be surveyed each year to measure their happiness in the workplace. Someone else has already done the hard part by working out the questions for Bhutan, though no doubt we'll have to add a few new ones for academic happiness (e.g. fairness in allocation – and appropriate recognition – of teaching and service roles, and availability of high quality coffee – or tea and biccies in my case).

An essential indicator of happiness will be the annual leave ratio (ALR). This is a simple calculation:

ALR = ALT/ALA

where ALT is the combined number of days of annual leave taken by all staff (unprompted by HR) for the previous year, ALA is the combined number of days of annual leave available to all staff each year (with a minimum of 20 days per staff member). Universities where all available annual leave is taken by staff would rate highest with an ALR of 1.

3. F-index

The F-index is about fairness. Despite published pay scales, men are paid more than women in the upper echelons of academia for doing the same job. Because, ‘loadings'. In my perfect world, to be eligible to participate in international rankings, universities would make public the average pay for men and women in leadership (professoriate and above) by posting the data on the front page of their website at the start of each year. The F-index is calculated as follows:

F = W/M

where W is the average salary for women in leadership positions and M is the average salary for men in leadership positions. Universities with the highest ratio (and thus the smallest gender pay gap) would rank highest on international rankings. It would be a fun experiment to see how long it would take for this indicator of university ranking to reverse the gender pay gap in academia. Looks like the University of Sydney will skyrocket on this measure: the VC indicated at the ‘Women at Sydney' event in late 2014 that remedying the gender pay disparity is a strategic objective for the university in 2015.

4. D-index

D is for diversity. Diversity is good. Differences in perspectives and methods of approaching problems lead to better outcomes. In Australia, our universities are populated by people of diverse culture, gender, age, socioeconomic status. Yet our leaders are mostly male and mostly white. The D-index measures how well the leadership teams at universities (professoriate and above) reflect the diversity of the broader university (all staff and students). For example, let's take gender. Plenty of studies show that more women in the workplace, especially more women in leadership positions, is not only the right thing to do, it's the smart thing to do because it's good for business. Yet gender equity in
leadership positions remains at dismally low levels (less than 20 per cent) across the board while male CEOs dig their heels in at quotas. To counter the entrenched system, the D-index for gender (Dg) will be evaluated: