Volume 198 - Issue 5
Risks of complaints and adverse disciplinary findings against international medical graduates in Victoria and Western Australia
Authors: David M Studdert, Matthew J Spittal and Katie Elkin
Med J Aust 2013; 198 (5): 258. || doi: 10.5694/mja12.11624
Published online: 18 March 2013
Published online: 18 March 2013
In reply: Nguyen and Nguyen question our finding that international medical graduates (IMGs) are at higher risk of attracting complaints than Australian-trained doctors (ATDs),1 based on recalculations using the descriptive statistics we reported. They infer, from the fact that IMGs accounted for 37% of doctors and 30% of total complaints, that IMGs were at lower risk of complaints.
Comparisons of simple proportions as above mislead for ...
In reply: Nguyen and Nguyen question our finding that international medical graduates (IMGs) are at higher risk of attracting complaints than Australian-trained doctors (ATDs),1 based on recalculations using the descriptive statistics we reported. They infer, from the fact that IMGs accounted for 37% of doctors and 30% of total complaints, that IMGs were at lower risk of complaints.
Comparisons of simple proportions as above mislead for a couple of reasons. First, doctors were observed for different periods of time — accounted for by conducting the analysis at the doctor-year level. Nguyen and Nguyen’s reanalysis rests on the incorrect assumption that duration of follow-up was comparable between the two groups. On average, observation periods for IMGs were significantly shorter than those for Australian-trained doctors. Secondly, comparisons of simple percentages do not account for confounders — characteristics that differ systematically between IMGs and Australian-trained doctors and are also associated with the probability of complaints. We adjusted for several of these characteristics with multivariate analysis. Nguyen and Nguyen’s calculations do not take them into account.
Nguyen and Nguyen’s other criticism is that we should have modelled the incidence of complaints as a count variable in Poisson regression, rather than as a binary variable in logistic regression. A count model is not clearly the superior method for this analysis, but it is a reasonable alternative. That is why we used one. We reran our analyses using negative binomial regression, a close cousin of Poisson regression. This sensitivity analysis produced very similar results, as reported in our original article and Appendix. Nguyen and Nguyen’s Poisson calculations may have produced different results for many reasons, including use of simple proportions rather than individual-level data, their erroneous assumption about equal follow-up time, and no adjustment for confounders.
Chaturvedi argues that there are practical and moral imperatives to better support IMGs in acclimatising to work and life in Australia. We agree. A recent parliamentary inquiry2 into overseas-trained doctors recognised this need, as Zubaran notes, and has called for immediate action.
Akre and Achhra object to our use of the term “high risk”. We used this term in a specific way — to denote groups of doctors with significantly higher odds of complaints and adverse findings — and it is accurate. Additionally, Akre and Achhra speculate that the higher risk of adverse findings observed for IMGs as a whole may have been driven by IMGs from English-speaking countries with relatively “severe” complaints against them. We did not report IMGs’ risks of adverse findings by country of training because the data were sparse (373 adverse findings across 21 specified countries), confidence intervals were wide, and the analysis was underpowered. However, the data do not support Akre and Achhra’s theory. We checked by running a multivariable model (with the same specifications as the complaints model) to predict adverse findings among doctors from the three English-speaking countries and three non-English-speaking countries (excluding Germany) with the most IMGs. Odds ratios (ORs) for doctors trained in the United Kingdom/Ireland (OR, 1.37; 95% CI, 0.66–2.85), New Zealand (OR, 0.69; 95% CI, 0.24–1.94) and South Africa (OR, 0.92; 95% CI, 0.33–2.59) were similar to those produced by the original risk-of-complaints model, as were ORs for IMGs from India (OR, 1.94; 95% CI, 1.17–3.23), Sri Lanka (OR, 1.93; 95% CI, 0.77–4.82) and Egypt (OR, 1.93; 95% CI, 0.87–4.25).
Breen notes that complaints to medical boards are uncommon, often do not result in adverse findings, and may stem from complainant-related biases. Our article addressed each of those points. Breen also suggests that reporting complaint risks by country of training was unnecessary and ethically questionable. This strong claim is difficult to square with the standard public health approach. Research and prevention activities are routinely targeted at certain groups of people. Consider, for example, efforts to combat youth suicide, falls among the elderly, obesity in low-income households, HIV infection among injecting drug-users and renal disease in Indigenous populations. Such targeting is widely accepted and occurs because foundational research has previously determined that these subgroups faced disproportionately high risks.
Why should complaints to medical boards be approached differently? Breen suggests that uncertainty about causal factors, particularly biases in complaint-filing behaviour and medical board adjudications, demands it. These and other factors (eg, lack of support, cultural misunderstandings) warrant exploration, and may help explain IMGs’ heightened risks. However, disentangling causal factors will take years and will require many studies. Publication of risk differences should not wait for this. Where would the impetus for such further research come from without initial identification of risk differences? Being less than candid about our findings would have done a disservice to IMGs and their patients.
Competing interests
No relevant disclosures.
References
- Elkin K, Spittal MJ, Studdert DM. Risks of complaints and adverse disciplinary findings against international medical graduates in Victoria and Western Australia. Med J Aust 2012; 197: 448-452. 0_CHDDCEDG
- House of Representatives Standing Committee on Health and Ageing. Lost in the labyrinth: report on the inquiry into registration processes and support for overseas trained doctors. Canberra: Parliament of the Commonwealth of Australia, 2012. http://www.aph.gov.au/Parliamentary_Business/Committees/House_ of_Representatives_Committees?url=haa/overseasdoctors/report.htm (accessed Feb 2013).