Abstract
This Chapter 10 builds on the lessons learned from traditional Doping Control to explore proposals for tackling the challenges that anti-doping faces and will continue to face under the 2015 WADC. Addressing these challenges requires that scientists and lawyers combine their expertise and their efforts to design paths to improvement (Sect. 10.1). These joint efforts should allow for a stronger qualitative approach to the gathering of evidence in Doping Control on the one hand (Sect. 10.2), and for greater refinement in the evaluation of the evidence gathered on the other hand (Sect. 10.3).
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Schamasch and Rabin 2012, p. 1698.
- 2.
See Sect. 6.2.2.3.1 above.
- 3.
See Sect. 10.2 below.
- 4.
See Sect. 10.3 below.
- 5.
Champod and Vuille 2011, p. 238: whereby an effective collaboration between forensic experts and lawyers supposes a clear definition of each others’ roles and realistic expectations regarding the nature of the expertise and its conclusions.
- 6.
Cooper 2012, p. 279: “When it comes to making practical and ethical policy there is a devil in the scientific detail (or “facts” if you like) that is absolutely required if we are to make informed moral and political choices”.
- 7.
See, for similar reflections in criminal proceedings, Champod and Vuille 2011, p. 227: “décalages existant entre les attentes des juristes et les réponses que peuvent apporter les experts aux questions d’expertise qui leur sont posées” [gap existing between the expectations of the lawyers and the answers that the experts can provide to the expertise questions they are asked (author’s translation)].
- 8.
Vuille 2011, p. 438.
- 9.
See Sect. 4.3.1.2 above; Champod and Vuille 2011, p. 228, speak of a “sentiment d’un dialogue de sourds” [feel of a dialogue of the deaf (author’s translation)].
- 10.
See Sect. 8.4.2 above.
- 11.
Marclay 2014, p. 8.
- 12.
Marclay 2014, p. 8.
- 13.
Jassanoff 1995, p. 221.
- 14.
Dwyer 2008, p. 139.
- 15.
See, in general, Sect. 6.3 above.
- 16.
Mazzoni et al. 2011, p. 611, also point to the fact that the Prohibited List represents a compromise between a “scientific, educational and legal tool and the format is the result of a compromise to suit such heterogeneity of purposes and audiences”.
- 17.
See Sect. 7.1.1.3.3 above.
- 18.
David 2013, p. 379: “if the focus of the Code can be tightened, it may be easier to maintain consistent, predictable outcomes”.
- 19.
Cooper 2012, p. 271: “the list could certainly be cleaned up to remove items that are not related to sports performance”.
- 20.
USADA Athlete Guide to the 2015 Prohibited List, http://www.usada.org/substances/prohibited-list/athlete-guide/ (accessed 15.05.2015); see the clenbuterol example, Sect. 9.1.2.2.3 above.
- 21.
See, denounced by Reymond J-F, “Le Code mondial antidopage n’a pas d’objectif clair” [the World Anti-Doping Code has no clear objective (author’s translation)], in Libération, http://www.liberation.fr/sports/2015/02/05/le-code-mondial-antidopage-n-a-pas-d-objectif-clair_1196505 (accessed 15.05.15).
- 22.
- 23.
For example, blood doping and the use of rhEPO, recognised as one of the most effective means of enhancing performance in endurance disciplines (see Zorzoli and Rossi 2010, Section Introduction; Lundby et al. 2012, p. 1307).
- 24.
See Sect. 10.2.1.1 below.
- 25.
See Sect. 7.1.2.1.4 above.
- 26.
See Sect. 7.1.2.1 above.
- 27.
See Sect. 7.2.1 above.
- 28.
See Sect. 6.3.2 above.
- 29.
See, in particular, Sect. 5.1.2.4.1 above.
- 30.
- 31.
See Sect. 5.1.2.4.1.1 above.
- 32.
See Sect. 7.1.2.1.2 above.
- 33.
See Sect. 6.3.3 above.
- 34.
See e.g., for Thresholds, Paul 2004a, p. 250.
- 35.
- 36.
In C-519/04, Meca-Medina & Majcen v. Commission, 18 July 2006, the Court of Justice of the European Union, without being completely clear on this issue, seems rather to consider that the burden of proof is on the Athlete who is challenging the Threshold to show a lack of proportionality in the level set by the ADO.
- 37.
See Sect. 10.3.1 below.
- 38.
See Sect. 5.1.1.3 above.
- 39.
See Sect. 5.1.1.1 above; see, as reiterated in the Q&A published by WADA on the revised 2015 WADC, https://www.wada-ama.org/en/questions-answers/2015-world-anti-doping-code#item-887 (accessed 15.05.15).
- 40.
Dvorak et al. 2014, Consensus Statement: “Successful anti-doping programmes of the future will embody high-quality, intelligent testing practices rather than high-quantity test volumes”.
- 41.
Rigozzi et al. 2013b, n° 38.
- 42.
See Sect. 5.1.1.3.1 above.
- 43.
Minutes WADA ExCo 11 May 2013, pp. 35 and 41.
- 44.
Botrè et al. 2014, p. 4: “no improvement [in sensitivity of the method] can counterbalance the wastefulness of a collection performed when the analytical target is no longer/not yet present in the biological sample”.
- 45.
Botrè et al. 2014, p. 4 et seq., and the Fig. 3 that shows the effectiveness of Testing as a function of the detection window for the relevant analytical method; for the ABP specifically, see Schumacher and d’Onofrio 2012, p. 984, stressing that the logistical costs for Sample collection and transport in the ABP exceed by far the analytical costs, so that the planning of and thinking behind each Testing is crucial.
- 46.
In this, the introduction of TD2014SSA represents a partial implementation of targeted analysis approaches, rather than full screening, as already advocated by anti-doping scientists for some years, see Saugy et al. 2011, p. 2.
- 47.
Indeed, TD2014SSA only determines the overall percentage of all Samples that must be tested for the substances covered by the Technical Document for the particular discipline, so that compliance with TD2014SSA can only be assessed a posteriori, based on the figures provided to WADA by the ADO at the end of a year.
- 48.
TDSSA2014, Supporting Document C, point 5.
- 49.
WADA ExCo Meeting 20 September 2014, p. 33.
- 50.
For further explanations of the process for setting the values in each sport, see WADA ExCo Meeting 20 September 2014, p. 27 et seq., especially p. 33.
- 51.
Dvorak et al. 2014; Saugy et al. 2014, p. 2.
- 52.
Saugy 2012, p. 661.
- 53.
Saugy et al. 2011, p. 2, whereby this “far more rational and efficient approach would allow to search for fewer substances that have been identified in the prevalence studies but in a larger number of athletes”.
- 54.
Saugy 2012, p. 649.
- 55.
Schamasch and Rabin 2012, p. 1695.
- 56.
See also Sect. 5.1.1.1.3 above.
- 57.
Lenz 2000, p. 61.
- 58.
Lenz 2000, p. 61. The suggestion has been made that in case of an equivalent potency of several fluids, the Athlete should be given a choice.
- 59.
For other substances, by contrast, the time of detection is longer in urine. The new steroidal module of the Athlete Biological Passport is thus based on urine.
- 60.
See Sect. 5.1.1.1.3 above; thus, blood sampling should not be conducted immediately prior to or during a Competition. In addition, the amount of blood taken must be limited to a quantity that cannot endanger the Athlete’s health. Blood sampling may be completely prohibited due to certain characteristics of the Athlete, for example haemophilia (see Lenz 2000, pp. 64 and 69).
- 61.
See Sect. 11.1.2 below.
- 62.
Saugy 2012, p. 660 with respect to substances prohibited In-Competition only; idem, ibidem, p. 659, with respect to substances with limited tolerated use.
- 63.
Cooper 2012, p. 267: “One of the problems with testing athletes is that good quality control is required from start to finish”, quoting the Lagat and Modahl cases as early cases of bad Sample handling resulting in bacterial degradation.
- 64.
Saugy et al. 2014, p. 4.
- 65.
Saugy 2012, p. 654.
- 66.
Saugy 2012, p. 653; see also Sect. 5.1.1.3.2 above.
- 67.
Botrè et al. 2014, p. 1/2; Saugy et al. 2014, p. 4, whereby the pre-analytical procedures “are certainly among the most sensitive and critical paths in the overall process” and transportation must be safe and rapid to ensure integrity of the Sample and avoid degradation.
- 68.
Botrè et al. 2014, p. 4; for an analysis of similar concerns in forensic sciences, Vuille et al. 2013, p. 1101.
- 69.
- 70.
See Sect. 5.1.1.3 above.
- 71.
Vuille et al. 2013, p. 1101: “les laboratoires sont tout désignés comme le lieu de tous les problèmes simplement parce que, de tout le processus forensique, c’est l’étape la plus facile à réglementer et à contrôler, et donc à blâmer en cas de problème” [laboratories are designated as the location for all problems simply because, within the overall forensic process, they are the easiest stage to regulate and to control, and thus to blame in case of problems (author’s translation)].
- 72.
See Sect. 5.3.2 above.
- 73.
Giraud et al. 2014, p. 333/334.
- 74.
- 75.
Marclay 2014, p. 9/10.
- 76.
Giraud et al. 2014, p. 334.
- 77.
Giraud et al. 2014, p. 332.
- 78.
See Sect. 5.1.1.3.3.1 above.
- 79.
Article 4.4.10 of the 2015 ISL and TD2014SSA newly provide that each laboratory is to make available through ADAMS the types of analyses that are within the scope of its technology and its related fees.
- 80.
See Sect. 5.1.1.2.2 above.
- 81.
Giraud et al. 2014, p. 344.
- 82.
See Sect. 5.1.1.3 above.
- 83.
See Sect. 2.3.1 above.
- 84.
Sottas 2010, p. 115.
- 85.
See Sect. 7.1.1.3 above.
- 86.
See Sect. 2.3.1 above.
- 87.
Giraud et al. 2014, p. 333/334.
- 88.
Saugy 2012, p. 658, highlights the shortcomings of qualitative analysis in urine for the examples of stimulants like pseudoephedrine, or glucocorticoids.
- 89.
See Sect. 10.3.1.2 below; Schumacher and d’Onofrio 2012, p. 979: “a conventional doping test sample that either contains a forbidden substance or not”.
- 90.
See Sect. 2.3.3.1 above.
- 91.
See Sect. 10.3.1.3 below; Sottas 2010, p. 123: “the increased complexity of both doping and anti-doping procedures increases the risk of logical fallacies if propositions and causal relationships between the pieces of evidence are not explicitly stated”.
- 92.
Article 5.4.4.2.1 of the ISL.
- 93.
See Sect. 7.1.2.2 above.
- 94.
See Sect. 7.3.4 above.
- 95.
See Sect. 6.2.3.3 above.
- 96.
See Sect. 2.3.3.2 above.
- 97.
Marclay et al. 2013, p. 135.
- 98.
Schamasch and Rabin 2012, p. 1997.
- 99.
Marclay 2014, p. 5.
- 100.
Sottas 2010, p. 123.
- 101.
Marclay 2014, p. 6; Dvorak et al. 2014, p. 2/4.
- 102.
Marclay 2014, p. 5; see for a “pilot experiment” implementing the suggested model, Harcourt et al. 2014.
- 103.
Sottas 2010, p. 123.
- 104.
Sottas 2010, p. 107.
- 105.
Cooper 2012, p. 265, for example, considers that a rate of false positives of 1:1,000 would be much too high to use the test as acceptable evidence in anti-doping proceedings, since the analysis is the “primary way of catching someone in the absence of any other evidence”, rather than just to “confirm whether someone is cheating”.
- 106.
Vuille 2014, p. 486; idem, ibidem, p. 488, stresses that DNA evidence nevertheless needs to be scrutinised for its own reliability, since multiple pieces of evidence mutually reinforce their probative value in the mind of the panels.
- 107.
For a case study, see Jan et al. 2011, especially p. 110 on the targeted use of DNA analysis, who stress that both urine and blood, the two matrices used for Doping Control, are compatible with DNA profiling.
- 108.
- 109.
For some first indications on this debate in the anti-doping movement, see Jan et al. 2011, p. 112/113.
- 110.
See e.g. Saugy 2012, p. 660, on adapting tools inspired from studies in public health to anti-doping, as a basis for more targeted Testing “campaigns” and analyses.
- 111.
See Sect. 6.2.2.2.3 above.
- 112.
See Sect. 3.1.4 above.
- 113.
For the assertion of a refusal to submit to Sample collection, and the question of whether the Athlete “left the station without having been told to do so in terms he could readily understand as being a formal injunction linked to a possible sanction”, which requires examination of conflicting testimonies, see CAS 2008/A/1551, WADA v. CONI, FIGC & Cherubin, para 64 et seq.
- 114.
Illustrations can be found for Attempted Use, or covering up and Complicity in an anti-doping rule violation, where the panel can only convict an Athlete or member of support personnel by deducing their intent from their acts. For Attempted Use, see CAS A4/2007, ASADA v. Wyper, para 36; for covering up or Complicity, see CAS 2008/A/1513, Hoch v. FIS & IOC, para 8.7.1 et seq.: “In the present case the Panel concludes that the Appellant knew of the anti-doping rule violation from a number of pieces of circumstantial evidence”; see also CAS 2007/A/1434 & 1435, IOC & WADA v. FIS & Pinter.
- 115.
We do not address here the particular situation of inferences being drawn against the Athlete for failure to appear at a hearing or refusal to answer the hearing panel’s questions (“silence in the face of the allegations”). Article 3.2.5 of the WADC explicitly authorises hearing panels to draw such inferences. This provision has been rightly questioned in legal literature in the light of the debate around the privilege against self-incrimination and the application of this principle in anti-doping proceedings (see Sect. 8.2.2.1.2 above).
- 116.
For more details, see Van Houtte 2009.
- 117.
David 2013, p. 138.
- 118.
David 2013, p. 138: the tribunal must be “comfortably satisfied that an inference which it is asked to draw follows from the given facts as something which is probably true”.
- 119.
See also Sect. 8.1.1.1 above.
- 120.
It is even questionable whether, as argued by McLaren 2012, p. 80, including an explicit mention of longitudinal profiling “directly in the WADC ensures the ABP is prima facie a ‘reliable means’”.
- 121.
This process was described in CAS 98/222, B. v. ITU, para 48 with respect to scientific assumptions underlying technical rules, as “the parties should share such responsibility and co-operate in the process of determination of the real cause”.
- 122.
CAS 2008/A/1718, IAAF v. All Russia Athletic Federation & Yegorova et al., para 179 et seq.
- 123.
CAS 2004/A/607, Boevski v. IWF, para 36.
- 124.
CAS 2008/A/1718, IAAF v. All Russia Athletic Federation & Yegorova et al., para 179: “The Panel finds that the natural, if not irresistible inference, is that the Athletes have somehow arranged to have the urine of third persons used in their out of competition testing”.
- 125.
CAS 2008/A/1718, IAAF v. All Russia Athletic Federation & Yegorova et al., para 179 et seq.; McLaren 2006a, p. 10.
- 126.
CAS 2004/A/607, Boevski v. IWF, 6 December 2004, para 34 et seq., in particular para 36: “Once the probability that manipulation occurred in this phase of the chain of custody becomes apparent then some explanation or plausible hypothesis that the athlete was not involved must be brought forward to refute the circumstantial evidence as to the probability of the manipulation either being carried out by the Appellant [the athlete] or with his consent and approval”; the panel also referred to the following paragraph in CAS 98/211, Smith-DeBruin, para 42: “Although invited to do so, Appellant’s counsel declined—and, in our view, was unable—to formulate any hypothesis that would point the finger at some such other person, whether identified or not. If and insofar as he invited us to consider in an abstract manner the possibility that either the Gs or some officer or employee of FINA were guilty of such manipulation, we utterly reject this suggestion”.
- 127.
See, in general, Sect. 3.1.2.2 above.
- 128.
CAS 2010/A/2384 & 2386, UCI & WADA v. Contador & RFEC, para 260 et seq.; for more details on the mechanisms, see Sect. 7.3.3.1.3 above.
- 129.
See Sect. 3.1.2.2 above.
- 130.
See Sect. 2.3.1.1 above on the causal chain of doping.
- 131.
See e.g. Senkel et al. 2014, p. 365.
- 132.
Even if we question the merits for making a distinction between direct and indirect evidence in anti-doping altogether (see Sect. 10.3.1.3 below), identifying “direct evidence” with “analytical evidence” is in any event inaccurate. The two situations may overlap to a large extent but are not necessarily identical by far. Non-analytical evidence may be indirect (e.g. a chain of emails between an Athlete and a “doping doctor” leading to the inference that the Athlete was using doping substances), but it may also be direct in the traditional legal meaning (e.g. the Athlete’s admission to having committed a violation). Inversely, analytical evidence may be direct evidence (e.g. detection of a Prohibited Substance in a Sample), but it may also be indirect (e.g. the detection of protease in a Sample leads to the conclusion that the Sample was tampered with to mask the Use of a Prohibited Substance).
- 133.
See for example, CAS A3/2007, ASADA v. Van Tienen, para 46; Rathgeber 2012, p. 1123; McLaren 2006a, p. 9.
- 134.
See e.g. Director of WADA Health, Medical & Research Committee, Valérie Fourneyron, Interview of 19 January 2015, https://www.wada-ama.org/en/media/news/2015-01/the-wada-interview-valerie-fourneyron (accessed 15.05.15); Berninger 2012, p. 101; David 2013, p. 161.
- 135.
CAS 2010/A/2235, UCI v. Valjavec & OC Slovenia, para 7: “direct detection methods aim to detect the doping agent itself”; David 2013, p. 161: “The ABP does not provide direct evidence of doping in the same way as a positive test, but requires the expert analysis of changes in an athlete’s profile which has been built up from the analysis of samples obtained over time”; Berninger 2012, p. 108, refers to a procedure based on indicia (“Indizienprozess”), which had to be strengthened by the WADA ABP Guidelines before ADOs would feel sufficiently secure to initiate anti-doping proceedings.
- 136.
See e.g. UCI Clean Sport website on the Athlete Biological Passport, http://www.uci.ch/clean-sport/the-athlete-biological-passport-abp/ (accessed 15.05.15); IAAF Athlete Advisory Note: Athlete Biological Passport 2013, p. 1, http://www.iaaf.org/about-iaaf/documents/anti-doping#athletes-guides-and-advisory-notes (accessed 12.07.14).
- 137.
See, in particular, Sect. 3.1.4.3 above.
- 138.
CAS 2010/A/2235, UCI v. Valjavec & OC Slovenia, para 7; UKAD Anti-Doping Panel Decision, British Cycling v. Tiernan-Locke, 15 July 2014, para 2/3.
- 139.
The Swiss Laboratory for Doping Analyses, and more specifically Dr Pierre-Edouard Sottas, developed the longitudinal approach and informatics tools underlying the ABP (see Lundby et al. 2012, p. 1311).
- 140.
Sottas 2010, pp. 116 and 121. To Sottas, the distinction between “direct” and “indirect” evidence of doping traditionally used in legal literature is meaningless: even for the most basic drug test, “an inferential logic must be used to evaluate the cause-to-effect relationship, such as how abuse of an exogenous steroid affects a mass spectrum obtained by GC-MS”.
- 141.
Sottas 2010, p. 116.
- 142.
Describing the same type of misjudgement when judges analyse forensic evidence, Vuille 2011, p. 166.
- 143.
On this shift in the legal definition of the violation, see Sect. 2.2.1.1 above.
- 144.
On the pharmacokinetic and pharmacodynamics processes, see Sect. 6.2.1.2 above.
- 145.
Thus, the ABP is also direct in the sense that it represents “direct evidence of a significant physiological effect” (Sottas 2010, p. 115).
- 146.
Sottas 2010, p. 121.
- 147.
Sottas 2010, p. 116.
- 148.
- 149.
Saugy 2012, p. 649.
- 150.
For an overview of these cases, see McLaren 2006b.
- 151.
Saugy et al. 2014, p. 5: “the T/E has been the first widely used indirect marker of doping with anabolic steroids, with a discrimination principle not based on the distinction between the exogenous substance and its endogenous counterpart, but rather on the effect induced by the intake of the exogenous substance on some selected biological marker”.
- 152.
Sottas et al. 2007.
- 153.
Vernec 2014, p. 2; Sottas 2010, p. 105, explains how the scientific basis for using T/E ratio as a sole evidence of doping turned out to be insufficiently strong.
- 154.
See Sect. 5.1.2.4 above.
- 155.
Giraud et al. 2014, p. 334.
- 156.
See Sect. 6.2.3.1.2 above.
- 157.
UKAD Anti-Doping Panel, British Cycling v. Tiernan-Locke, 15 July 2014, para 12; see also Schumacher and d’Onofrio 2012, p. 979: “a priori fact independent from the involvement of the expert in the case”.
- 158.
Note that these approaches have been developed to target blood doping, a domain in which the possibility of encountering Athletes who could make a plausible defense of contamination or other inadvertence is rather remote, another factor that may have promoted the assimilation of “violation” with “doping” in the scientists minds.
- 159.
Sottas 2010, p. 115.
- 160.
Sottas et al. 2008a, p. 206, on Bayesian approaches for blood doping markers: “Because of the high number of variables, this framework may appear complex at first sight. However, we believe that the testing procedure gains in clarity and fairness by making all these variables explicit”.
- 161.
See the assessment in Sect. 7.3.4 above with respect to the presumption of Fault, and for possible tools to enhance this assessment, Sect. 10.3.2.2.3 below.
- 162.
See Sect. 3.1.4.1 above.
- 163.
Durston 2011, p. 122/123; Murphy and Glover 2011, p. 3: “A judicial trial is not a search to ascertain the ultimate truth of the past events inquired into, but to establish that a version of what occurred has an acceptable probability of being correct. It is in the nature of human experience that it is impossible to ascertain the truth of past events with absolute certainty”; see also Anderson et al. 2005, p. 246, who list five reasons for which legal conclusions based on evidence are necessarily probabilistic: (i) the available evidence is always incomplete, (ii) evidence is commonly inconclusive, (iii) evidence is often ambiguous, (iv.) bodies of evidence are commonly dissonant, and (v.) evidence comes from sources with different gradations of credibility.
- 164.
For a similar view from a Swiss perspective, Berger-Steiner 2008, p. 263/264.
- 165.
Taroni and Biedermann 2010, p. 340; Vuille 2011, p. 136. As these authors stress, this in no way means that these probabilities are “arbitrary” or cannot be deducted from investigations and empirical research, but simply that the probabilities are not “frequentist” or “objective”, since events that are of interest in the judicial process are usually unique so that their recurrence cannot be tried out through repetition in experimental conditions; see, however, for an overview regarding the current of doctrine based on objective probabilities, Berger-Steiner 2008, p. 152 et seq., who adds that in most cases the empirical data does not exist. On the other hand, Vuille 2011, p. 168, stresses that probabilities used in the judicial process do not amount to a denial that a fact either happened or did not happen in reality, they only express degrees of belief as to whether something happened or not (i.e. they are epistemological and not ontological).
- 166.
Kaufmann 2009, p. 190: “Nur bei wenigen naturwissenschaftlichen Beweisen (DNA-Vergleiche, Messung der Blutalkoholkonzentration) kann die Wahrscheinlichkeit in Form einer Prozentangabe angegeben werden” [there are only few proof in natural sciences (DNA-comparisons, measurement of the alcohool concentration in blood) for which the likelihood can be given as a percentage, (author’s translation)].
- 167.
Murphy and Glover 2011, p. 3.
- 168.
Durston 2011, p. 122/123; a notable exception being the “balance of probability” standard, which is commonly defined by legal hearing panels as exceeding the 50 % likelihood threshold (see Sect. 7.3.3.1.3 above); Kaufmann 2009, p. 190: “Wird in anderem Zusammenhang gesagt: ‘Das Gericht ist zu 95 % überzeugt, dass der Angeklagte das Opfer getötet hat’, ist dies nicht wörtlich zu verstehen, sondern vielmehr als Versuch, den Grad der Ueberzeugung zu messen” [If, in a different context, it is said: “The court is satisfied to 95 % that the accused killed the victim, this is not to be understood literally, but as an attempt to measure the degree of persuasion” (author’s translation)].
- 169.
Taroni and Aitken 1998, p. 293/294; for an overview of attempts to formalise the evaluation of evidence among Swedish scholars, see Lindell 2004, p. 428 et seq.
- 170.
For an overview in the common law context, see Anderson et al. 2005, p. 247.
- 171.
For an overview, see Dwyer 2008, p. 53/54.
- 172.
Taroni and Biedermann 2010, p. 339.
- 173.
See Chapter 11 below for the Athlete Biological Passport.
- 174.
See Sect. 9.3.2 above.
- 175.
See Sect. 2.3.1.1 above.
- 176.
See Sect. 10.3.2.2.3 below.
- 177.
For a—rare—presentation of the Bayes Theorem in a Swiss law study, see Berger-Steiner 2008, p. 264, arguing against its general application in the evaluation of evidence.
- 178.
Saugy 2012, p. 653.
- 179.
Taroni and Aitken 1998, p. 293/294; Vuille et al. 2013, p. 1105, footnote 22, whereby “elle est cependant généralement admise par la grande majorité de la littérature juridique et scientifique comme étant le meilleur modèle existant” [nevertheless it is generally accepted in the large majority of the legal and scientific literature as the best existing model, (author’s translation)], with further references; Taroni and Biedermann 2010, p. 342, further note that the theorem itself is universally recognised as being mathematically correct.
- 180.
See the description of the defensive reaction of US courts to the introduction of Bayesian tools in jury trial in Aitken and Taroni 2008, p. 182/183.
- 181.
Vuille 2011, p. 171.
- 182.
See Aitken and Taroni 2008, p. 183: “the probabilistic foundation of evidence is a fact of life”.
- 183.
Vuille 2011, p. 172. [“In our view, the resistance some lawyers show against the use of probabilities for evaluating evidence is all the more curious since none of them would reject the idea that the judge must reason in a logical manner; and probabilities are nothing but formalised logic”, (author’s translation)].
- 184.
Taroni and Aitken 1998, p. 304.
- 185.
Aitken and Taroni 2004, p. 78.
- 186.
In which case the fallacy is known as the “prosecutor’s fallacy”.
- 187.
See Sect. 10.3.2.3.3 below.
- 188.
See Sect. 10.3.2.3 below.
- 189.
Vuille 2011, p. 80, whereby any value that the expert attributes to a piece of evidence in probabilistic terms can never be converted directly into the percentage necessary to assess whether the standard of proof is reached.
- 190.
Champod and Vuille 2011, p. 234, characterise the debates around quantifying the concepts of “reasonable doubt” or “intime conviction” as “legal heresy”, in that intime conviction can only result from the overall weighing of the evidence on the record and thus can only be the task of the judge.
- 191.
Taroni and Biedermann 2010, p. 341.
- 192.
Vuille 2011, p. 169.
- 193.
Sottas 2010, p. 107.
- 194.
To a certain extent, since the 2015 WADC places greater emphasis on Target Testing based on intelligence. It is still true in the sense that there is currently no model for integrating into the subsequent proceedings resulting from the Adverse Analytical Finding considerations that the positive finding was the consequence of Target Testing, or of purely random Testing.
- 195.
The situation is thus comparable to the situation of a DNA match in which the suspect is determined merely based on his presence in a DNA profile database, for which all the a priori probabilities of the individuals present in the database are equal (see Vuille 2011, p. 218). The rate of false positive and the size of the relevant population, as well as the prevalence of doping, must be given particular consideration, since there are no a priori elements in the file to corroborate the fact that the match (or here, the positive analytical finding), is a “real” as opposed to a “false” positive finding.
- 196.
Sottas 2010.
- 197.
See Sect. 8.1.2.2.1 above.
- 198.
See Sect. 10.3.1.3 above.
- 199.
See Sect. 10.3.1.2 above.
- 200.
See Sect. 8.1.1.1 above.
- 201.
Anderson et al. 2005, p. 56.
- 202.
Also referred to here as “false positive” and “specificity” issues (as the WADC regime does for the decision limits in rhGH, see WADA hGH Guidelines, section 7), even though in a narrow meaning these issues no longer pertain to the method itself, but to the interpretation of the raw data obtained from the analysis to decide whether the data is indicative of exogenous hGH or not.
- 203.
Vuille et al. 2013. p. 1098, whereby the false positive should be taken into account when receiving a likelihood ratio from the expert, so that the likelihood ratio should in reality have two denominators (i.e. the likelihood of a fortuitous match + the likelihood of reporting a non-match as a match).
- 204.
See Sect. 5.2.4.2 above.
- 205.
On other challenges often imprecisely described as “false positives”, see Sect. 5.2.4.1 above.
- 206.
CAS 2005/A/831, IAAF v. Hellebuyck, para 7.2.4.2, with numerous references to earlier CAS awards re rhEPO; CAS 2011/A/2566, Veerpalu v. FIS, para 203.
- 207.
See Sect. 5.2.4.2 above.
- 208.
On the purpose-tied character of reliability, see Sect. 4.3.2.1.1 above.
- 209.
Fischer K and Berry D, Statisticians Introduce Science to International Doping Agency: The Andrus Veerpalu Case, Chance Magazine, online publication: http://chance.amstat.org/2014/09/doping/ (accessed 31.05.2015), section “Discussion”.
- 210.
Sottas 2010, p. 108/109.
- 211.
Sottas et al. 2008b, p. 166.
- 212.
Sottas 2010, p. 113.
- 213.
See the exchanges of view in the journal Nature, between Berry 2008, p. 692/593, and Sottas et al. 2008b, p. 166.
- 214.
See Sect. 6.3.1 above.
- 215.
Thus, the CAS panel in the unpublished case USADA v. Bergman is reported to have declared that its role “required it to be satisfied that the risk of a false positive for an athlete was at an acceptably low level to establish the doping offence” (see McLaren 2006b, n° 2.b). This would suppose that the hearing panel be given tools to “cast a bridge” between the false positive rate and the standard of proof.
- 216.
See Sect. 10.3.2.2.3 below, on the Positive Predictive Value as a possible tool for giving a better meaning to the rate of false positive.
- 217.
CAS 2011/A/2566, Veerpalu v. FIS, para 204.
- 218.
CAS 2005/A/831, IAAF v. Hellebuyck, para 7.2.4.2.
- 219.
CAS 2011/A/2566, Veerpalu v. FIS, para 203.
- 220.
See e.g. CAS 2005/A/831, IAAF v. Hellebuyck, para 7.2.4.2 for the series of rhEPO cases previously assessed by CAS panels; for example, in CAS 2003/A/452, IAAF v. MAR & Boulami, para 30, the argument by the Athlete’s expert: “suggests a much higher risk of false positives in the EPO test than originally thought”.
- 221.
CAS 2011/A/2566, Veerpalu v. FIS, para 203, footnote 249; see, inaccurately, CAS 2014/A/3488, WADA v. Lallukka, para 4, citing the findings of the Veerpalu v. FIS award as “found that the risk of false positive tests was too high”.
- 222.
The absence of real control is apparent in the language used in the CAS awards. For example, in CAS 2003/A/452, IAAF v. MAR & Boulami, para 33, whereby the ADO had demonstrated that the applicable decision limits “largely eliminates the risk of false positives”; para 40 “each laboratory determined that the risk of false positives would be virtually nonexistent”.
- 223.
In forensic sciences, the discovery of false positives is often due to “fruit du hasard”, in particular because the results obtained are simply improbable and the analysis is conducted anew, see Vuille et al. 2013, p. 1102.
- 224.
See Sect. 10.2.1.1 above.
- 225.
On the use of additional evidence in violations under Article 2.1 of the WADC, see Sect. 8.1.2.2.1 above.
- 226.
The challenge being to identify the population of interest, see Sect. 10.3.2.3.3 below.
- 227.
On the importance of considering how the rate of false positive affects the estimate probative value in forensics, Taroni and Biedermann 2010, p. 354; Vuille 2011, p. 164, highlights the importance of avoiding the “fallacy of the false positive”, which is one illustration of the inversion fallacy. This fallacy consists in assuming, when being informed of the rate of false positives produced by an analytical test (e.g. 1:10,000), that the particular positive analytical finding at stake thus has 1:9,999 chances of being a “real” positive. This is inaccurate, since the actual odds depend also on the a priori probability, in particular the prevalence of doping in sport.
- 228.
Sottas et al. 2008a, p. 196.
- 229.
Sottas et al. 2008a, p. 196.
- 230.
Sottas 2010, p. 110.
- 231.
See on these concepts, Sect. 5.2.4.2.1 above.
- 232.
Sottas et al. 2008a, p. 196: “the size of the number of tests elevates the likelihood of finding a match by pure chance alone”; McLaren 2012, p. 94.
- 233.
Putting aside here the challenges inherent in gathering data about prohibited behaviors (see e.g. Sottas et al. 2011, p. 763).
- 234.
See Sect. 10.3.1.2 above on the ambiguities surrounding direct versus indirect evidence of doping.
- 235.
See Sect. 2.3.1.1 above.
- 236.
Sottas et al. 2011, p. 762.
- 237.
By definition, only Athletes who know that they ingested a Prohibited Substance or Used a Prohibited Method can state so. This is always assuming that Athlete respond truthfully, something that especially elite Athletes may be hesitant to do (see Sottas et al. 2011, p. 763).
- 238.
The two means of study would only provide equivalent results for substances or methods that can almost by definition only reflect a deliberate ingestion or Use.
- 239.
On these distinctions, see Sect. 5.2.4 above.
- 240.
This would correspond to an equation in which the SP/SE data would only express the rate of false positive inherent in the design of the method, and the PD would be calculated based on analytical comparisons of expected values.
- 241.
It does not then strictly speaking express the probability that “an Athlete doped”. See, imprecisely in our view, Sottas 2010, p. 110.
- 242.
In particular, the estimate SP/SE could incorporate, in addition, the likelihood of a procedural defect having occurred during Doping Control, the risk of ingestion through contamination beyond the Athlete’s control, or even the risk of ingestion through an excusable inadvertence on part of the Athlete, always with respect to the Prohibited Substance or Prohibited Method at stake.
- 243.
Note that in any event, the PPV always indicates the direction P(E/H), thus never directly the P(H/E) that a hearing panel is ultimately looking for.
- 244.
CAS 2014/A/3475, Van Snick v. FIJ, para 106: “Compte tenu de la très faible quantité de métabolites de cocaïne détectée et de la chronologie vraisemblable des événements, la Formation arbitrale estime qu’elle peut se montrer un peu plus souple quant à l’appréciation de la preuve apportée par l’athlète relative à la façon dont la substance interdite s’est retrouvée dans son organisme” [Given the very low quantity of cocaine Metabolites detected and given the likely chronology of the events, the arbitration panel considers that it can show itself a bit more flexible in the evaluation of the evidence produced by the Athlete with respect to the manner in which the prohibited substance came into his organism, (author’s translation)].
- 245.
Sottas 2010, p. 109, only assesses the Positive Predictive Value assuming two hypotheses, i.e. the Athlete “doped” or the Athlete was “clean”. In practice, however, one would need to determine whether intermediate situations (e.g. a Sample contamination or ingestion of clenbuterol through contaminated meat) is to be included into Hdoping or into Hclean. This, again, shows the importance of having both scientific and legal expertise at hand when developing tools for legal purposes in anti-doping.
- 246.
Sottas 2010, p. 123.
- 247.
Taroni and Biedermann 2010, p. 356, insist that even the fact that the rate of false positives be unknown does not change the merits of taking it into account in the probabilistic reasoning; for an example applied to anti-doping, see Sottas 2010, p. 124.
- 248.
Vuille et al. 2013, p. 1108/1109: “l’évaluation de la valeur probante ne devrait pas être vue comme une prérogative de l’expert, mais comme une opportunité pour toutes les parties impliquées d’articuler leurs hypothèses quant au cas investigué” [the evaluation of the probative value should not be seen as a prerogative of the expert, but as an opportunity for all parties involved to articulate their hypotheses with respect to the matter under investigation, (author’s translation)].
- 249.
Taroni and Biedermann 2010, p. 343.
- 250.
See Sect. 2.3.1.3 above.
- 251.
Vuille et al. 2013, p. 1105, footnote 22.
- 252.
Taroni and Aitken 1998, p. 294.
- 253.
The calculation can also be made in terms of “odds”, i.e. the ratio between the probability for the hypothesis of interest and its complement hypothesis, in which case the formula is: a posteriori odds = a priori odds x likelihood ratio (see Aitken and Taroni 2008, pp. 189 and 192).
- 254.
The terms P(H1/E,I), express the fact that the new evidence [E] comes on top of the pre-existing information [I], but the term [I] will be left out for the sake of simplicity throughout the presentation in this book.
- 255.
Taroni and Biedermann 2010, p. 344.
- 256.
Champod and Vuille 2011, p. 236: “la force d’un élément matériel ne peut s’exprimer que de manière relative en considérant les résultats techniques sous l’angle d’au moins deux propositions (comme ici la thèse alléguée par l’autorité de poursuite et celle proposée par la défense). Un expert ne peut donc pas orienter une thèse si aucune alternative ne lui est fournie” [the strength of a material element can only express itself in a relative manner, by considering the technical results from the perspective of at least two propositions (such as here the thesis alleged by the prosecuting authority and the one proposed by the defence). Thus, an expert is not in a position to orientate a thesis if he is not provided with a set of alternatives, (author’s translation)].
- 257.
Aitken and Taroni 2008, p. 194.
- 258.
Vuille et al. 2013, p. 1099: “résume la valeur probante d’un indice en relation avec deux hypothèses de travail” [summarises the probative value of an indicium with respect to two working hypotheses, (author’s translation)]; in continental European literature, the likelihood ratio is also referred to as “abstract” probative value, while the probability a posteriori is described as “concrete” or “actual” probative value.
- 259.
Vuille 2011, p. 146/147; Vuille et al. 2013, p. 1097, footnote 6.
- 260.
Aitken and Taroni 2008, p. 193.
- 261.
Taroni and Mangin 1999, p. 11; a different situation seems to prevail for paternity claims, in which the expert directly provides the judge with the probability a posteriori that an individual is the father of the child, which requires the expert to determine him- or herself a priori probabilities without the benefit of the file, a situation strongly criticised by leading commentators (see Vuille 2011, p. 96, and the references cited).
- 262.
Aitken and Taroni 2004, p. 4.
- 263.
Taroni and Aitken 1998, p. 300; Taroni and Biedermann 2010, p. 344.
- 264.
McLaren 2012, p. 84.
- 265.
Vuille et al. 2013, p. 1097: “un rapport de vraisemblance élevé en faveur d’une hypothèse ne signifie pas encore que l’hypothèse elle-même est probable ou très probable; ce passage de la probabilité d’une concordance à la probabilité d’une hypothèse nécessite encore la prise en considération des autres éléments du cas d’espèce” [a very high likelihood ratio in favour of one hypothesis does not mean that the hypothesis itself is likely or very likely; this transition from the likelihood of a concordance to the probability of a hypothesis supposes, in addition, that the other elements of the case be taken into account, (author’s translation)].
- 266.
Taroni and Aitken 1998, p. 302; Vuille 2011, p. 170, considers that this solution is a better one than to adopt a random (generally 0.5) a priori probability in the absence of case-relevant information.
- 267.
For the sake of the presentation, the fact that there are typically two hypotheses being confronted is omitted here.
- 268.
Champod and Vuille 2011, p. 234; see, in general, Sect. 4.3.1.2.3 above.
- 269.
Vuille 2011, p. 187/188.
- 270.
Vuille 2011, p. 187/188.
- 271.
In criminal proceedings, the relevant issue might be: did X commit the rape he is accused of?
- 272.
In determining the probability that the DNA collected on the victim is X’ DNA, which can be based mostly on statistical information and the DNA analysis.
- 273.
For the activity level (“what is the probability of the evidence assuming X had sexual intercourse with the victim?”), the expert must be provided additional information from the file, e.g. that would point at the possibility that the sperm of the accused was found on the location of the offence for a reason other than having intercourse with the victim.
- 274.
Wilson 2009, p. 127.
- 275.
Sottas 2010, p. 122.
- 276.
Taroni and Aitken 1998, p. 295.
- 277.
Taroni and Aitken 1998, p. 300.
- 278.
Anderson et al. 2005, p. 460.
- 279.
Vuille 2011, p. 165, footnote 425; idem, ibidem, p. 148, points out that the use of a likelihood ratio or bayesian scheme for certain pieces of evidence by no means implies that the entire file must be submitted to the same scheme, and in particular, does not imply that the final legal decision be reached mathematically.
- 280.
We are not trying to make a determination here on the general usefulness of Bayes’ Theorem for the evaluation of the evidence in the judicial process; see Berger-Steiner 2008, p. 265, questioning the usefulness of the tool, in particular because of the frequent lack of data on the probability a priori and the difficulty to render the “vielschichtigen Lebenssachverhalte” [multifaceted life circumstances] of the civil proceedings.
- 281.
Aitken and Taroni 2008, p. 203; Coquoz 2006, p. 250.
- 282.
Taroni and Biedermann 2010, p. 3: “L’approche bayésienne pour l’analyse de données scientifiques permet une séparation naturelle et logique entre l’appréciation d’un problème donné avant l’acquisition de nouvelles informations et le moment de cette acquisition. Ces deux moments sont clairement identifiés comme étant importants dans la formation d’une vision complète d’un problème et offrent un moyen de combiner ces deux types d’information” [The bayesian approach for the analysis of scientific data allows for a natural and logical separation between the appreciation of a given problem before acquiring new information and the moment of this acquisition. These two moments are clearly identified as being important as part of forming a complete vision of a problem and offer a tool to combine these two types of information, (author’s translation)].
- 283.
Coquoz 2006, p. 251; Vuille 2011, p. 168, stressing that the likelihood ratio is ultimately a vector of logic.
- 284.
Vuille 2011, p. 171: “En revanche, et c’est peut-être là que le débat est légèrement dépassé, la complexité toujours plus grande des preuves scientifiques, et le recours toujours plus fréquents que la justice y fait, nécessite qu’on s’arme d’instruments adaptés pour garantir au justiciable que sa cause sera traitée rationnellement. Or, évaluer une preuve scientifique de façon rationnelle nécessite certainement qu’on l’appréhende avec plus que le simple ‘flair’ qui fait les bons enquêteurs” [By contrast, and this is probably where the debate is slightly obsolete, the ever greater complexity of scientific evidence, and the ever more frequent resorting to such evidence in the judicial process, requires one to arm oneself with instruments adapted to guarantee to the person subjected to the jurisdiction of the courts that his or her case will be treated in a rational manner. Indeed, the evaluation of scientific evidence in a rational manner requires undoubtedly that one should approach it with more than the mere “flair” that makes good investigators, (author’s translation)].
- 285.
See, in particular, Sect. 10.3.1 above.
- 286.
Coquoz 2006, p. 247.
- 287.
This process calls on the so-called “inductive logic”, see Anderson et al. 2005, p. 56/57.
- 288.
Champod and Vuille 2011, p. 237: an expert can never express him- or herself on the plausibility of a hypothesis considered separately and in an isolated manner (e.g. it is “possible”, “highly likely”, “cannot be excluded” that something happened).
- 289.
Anderson et al. 2005, p. 90; Champod and Vuille 2011, p. 236; in criminal law, when the defence is not putting forward a hypothesis, the assumption is usually that the hypothesis is the complement of the position of the prosecution, whatever this position (see Aitken and Taroni 2008, p. 187).
- 290.
On the manner in which this difficulty has led to a standardisation of evidence in traditional Doping Control, see Sottas 2010 and Sect. 7.3.4 above.
- 291.
Anderson et al. 2005, pp. 57 and 90.
- 292.
Anderson et al. 2005, p. 264.
- 293.
For anti-doping, see Sottas 2010, p. 121.
- 294.
Aitken and Taroni 2004, p. 101.
- 295.
See Sect. 10.3.2.3.1 above.
- 296.
Aitken and Taroni 2004, p. 80; Champod and Vuille 2011, p. 233, describes this recurring error both on the part of the scientists and on the part of the lawyers as a “piège de l’intuition, qui consiste à confondre les probabilités associées aux effets avec les probabilités des causes” [pitfalls of the intuitive approach, which consists in confusing the probabilities related to the effects with the probabilities related to the causes, (author’s translation)]; or, in mathematical terms, confuse P(E/H) with P(H/E); for a case study explained to legal practitioners, see Aitken and Taroni 2008, pp. 184-187.
- 297.
Schumacher and d’Onofrio 2012, p. 980, on the importance to “highlight the direction of the assessment of evidence” as a standard to be applied to the evaluation of evidence in the Athlete Biological Passport.
- 298.
Taroni and Biedermann 2010, p. 345.
- 299.
Taroni and Biedermann 2010, p. 343. In fact, the two probabilities can be converted into each other through the following formula: P(H/E) = P(E/H) x P(H)/P(E). This formula shows that the two probabilities are only equal when P(H) = P(E).
- 300.
An illustration of this fallacy can be found in Decision of the Arbitration Committee of the Czech Olympic Committee, Kreuziger v. Czech Cycling Federation, 22 September 2014. The fallacy could mean, in particular, concluding from the fact that an analytical test produces a rate of false positives of 1:100, to the fact that the odds that an Athlete who actually tested positive are 99:100 that the Athlete doped. In reality, these odds depend on other factors, primarily the prevalence of doping in the relevant Athlete population. The likelihood of having struck a false positive is much higher, in relative terms, if the prevalence is 2 %, than if the prevalence is 80 %, for example.
- 301.
Aitken and Taroni 2004, p. 82.
- 302.
For another purpose, i.e. a more targeted Sample analysis menu, see Sect. 10.2.1.1 above; for the use for the Positive Predictive Value, see Sect. 10.3.2.2.3 above.
- 303.
Sottas 2010, p. 122.
- 304.
On the need to clearly circumscribe the population of interest when evaluating the prevalence of doping, given the regional variability as well as variability in different sports, Saugy 2012, p. 654.
- 305.
On the difficulties in determining the relevant population in the forensic context, Vuille 2011, p. 141; Aitken and Taroni 2008, p. 189, point at the fact that the relevant population often depends on the arguments of the defence as much as on the arguments of the prosecution; on possible tools for studying prevalence, see Sect. 10.2.2.3 above.
- 306.
Sottas 2010, p. 105/106.
- 307.
Sottas 2010, p. 106.
- 308.
Obviously, the discussions would then need to revolve around what “clean” means, both scientifically and legally speaking, see Sect. 10.3.2.2.3 above.
- 309.
See Sect. 7.3.3.2.2.1 above.
- 310.
Putting aside all issues of Sample contamination or procedural defects, the only other information that the Adverse Analytical Finding implicitly gives for framing a hypothesis is the detection window of the test (see Sottas 2010, p. 113).
- 311.
Recognised as an expression of evidence necessity (“Beweisnotstand”) in CAS 2010/A/2384 & 2386, UCI & WADA v. Contador & RFEC, see Sect. 7.3.3.1 above.
- 312.
Sottas 2010, p. 109, concludes that, precisely due to the Strict Liability rule in the WADC regime, it is important “to have solid knowledge of the evidentiary value of an anti-doping test”.
- 313.
See for a detailed presentation, Sottas 2010, p. 103 et seq.
- 314.
See e.g. CAS 2010/A/2308 & 2011/A/2335, Pellizotti v. CONI & UCI, UCI v. Pellizotti, FCI, CONI, para 49; Verbiest 2010, p. 134.
- 315.
Decision of the Arbitration Committee of the Czech Olympic Committee, Kreuziger v. Czech Cycling Federation, 22 September 2014, para 6.4.
- 316.
For a comparative table of the different forms in which numerical scales (probabilities, percentage, odds) may correlate with verbal scales, see Anderson et al. 2005, p. 230; for other examples, see Aitken and Taroni 2008, p. 202.
- 317.
Taroni and Aitken 1998, p. 298, note 30.
- 318.
Aitken and Taroni 2004, p. 107.
- 319.
Anderson et al. 2005, p. 247.
- 320.
Aitken and Taroni 2008, p. 195.
- 321.
Vuille 2011, p. 179; see Sect. 10.3.2.3.1 above.
- 322.
For an example of such erroneous conclusion (fortunately only obiter dictum), Decision of the Arbitration Committee of the Czech Olympic Committee, Kreuziger v. Czech Cycling Federation, 22 September 2014, para 6.4; Vuille 2011, p. 179, notes that there is a particular risk if the probability a priori was low, in which case erroneously using the likelihood ratio as the probability a posteriori will be extremely detrimental to the accused.
- 323.
Some suggestions have been formulated in the scientific anti-doping community, in the context of the Athlete Biological Passport, see Sect. 11.4.2.2.3 below.
- 324.
See Chap. 11 below.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2016 T.M.C. Asser Press and the author
About this chapter
Cite this chapter
Viret, M. (2016). Paths to Improvement Through New Approaches. In: Evidence in Anti-Doping at the Intersection of Science & Law. ASSER International Sports Law Series. T.M.C. Asser Press, The Hague. https://doi.org/10.1007/978-94-6265-084-8_10
Download citation
DOI: https://doi.org/10.1007/978-94-6265-084-8_10
Published:
Publisher Name: T.M.C. Asser Press, The Hague
Print ISBN: 978-94-6265-083-1
Online ISBN: 978-94-6265-084-8
eBook Packages: Law and CriminologyLaw and Criminology (R0)
