Writing 1 PSY assignment

Jean2979

article2dopinginsportsworldissue.pdf

Home >Psychology homework help >Writing 1 PSY assignment

RESEARCH REPORT

doi:10.1111/j.1360-0443.2006.01568.x

Addiction,

101

, 1640 – 1644

Blackwell Publishing Ltd

Oxford, UK

ADDAddiction

2006

101

1116401644

Original Article

Doping in fitness sports Perikles Simon et al.

Correspondence to:

Heiko Striegel, University Medical Clinic, Department of Sports Medicine, University of Tuebingen, Silcherstraße 5, 72076 Tübingen, Germany. E-mail: heiko.striegel@uni-tuebingen.de

Submitted 7 December 2005; initial review completed 9 January 2006; final version accepted 22 May 2006

RESEARCH REPORT

Doping in fitness sports: estimated number of unreported cases and individual probability of doping

Perikles Simon

, Heiko Striegel

, Fabian Aust

, Klaus Dietz

& Rolf Ulrich

University of Tuebingen, Department of Sports Medicine, General Internal Medicine V, Tuebingen, Germany,

University of Tuebingen, Department of Cognitive and

Biological Psychology, Tuebingen, Germany

and University of Tuebingen, Department of Medical Biometry, Tuebingen, Germany

ABSTRACT

Aims

Recent studies have suggested that the use of doping substances and particularly of anabolic androgenic steroids (AAS) is often practised by fitness centre visitors. These studies employed direct interview techniques and questionnaires to assess the estimated number of unreported cases of doping. Because people hesitate to provide com- promising information about themselves, these techniques are subject to response errors. In this study we applied an alternative interview technique to assess more accurately unreported cases of doping in fitness centres.

Design and participants

The present investigation employed the randomized response technique (RRT) to reduce response errors. A cohort of 500 people from 49 fitness centres participated in this study.

Finding

The RRT revealed a high pre- valence of doping (12.5%). In addition, and most importantly, the present RRT study revealed an alarmingly high pre- valence of illicit drug use, specifically of cocaine use, that has been severely underestimated by previous studies.

Conclusions

The RRT confirmed previously estimated rates of AAS use assessed by direct interview techniques and voluntary questionnaires, but uncovered a much higher usage rate of illicit drugs among fitness centre visitors. This outcome enabled us to construct a ‘probability’ rating for the use of doping substances in fitness centre visitors. Given its high prevalence and the predominant use of AAS, doping among fitness centre visitors is an issue of extreme rele- vance for the health care system. Our study may help to characterize further doping substance users and to develop and apply prevention and intervention programmes specifically to individuals at high risk.

Keywords

Doping, fitness sports, randomized response technique, prevention.

INTRODUCTION

Doping substances are used by increasing groups of the younger population for life-style purposes [1–4]. The most commonly abused substance class is anabolic androgenic steroids (AAS), despite its relatively high short- and long-term risks [5–8]. Previous studies in western Europe and the United States suggested a preva- lence of AAS use among male fitness centre visitors of about 5–10% [4,9]. This is an alarming figure given that, according to official figures of the German fitness centre organization, in 2004 nearly 5 million people visited fitness centres (http://www.dssv.de). It has been extrapolated that male fitness centre visitors contribute to as many as 1 million AAS users in the United States [1,10,11].

Prevention and intervention programmes against doping may therefore benefit from specifically targeting fitness centre visitors. It is important to isolate the sub- group of doping substance users, as fitness centre visitors are a heterogeneous population with regard to many anthropometric, social and training parameters [12]. Accordingly, a more precise characterization of the dop- ing substance users with regard to the above-mentioned basic parameters will therefore be helpful in planning and initializing anti-doping programmes.

The results of our previous study suggest that doping substance users in fitness centres constitute a unique group, distinct from the group of illicit drug users concerning anthropometrics, social indicators and training data [12]. As in the case of other survey studies that have employed direct interview techniques,

http://www.dssv.de

Doping in fitness sports

1641

Addiction,

101

, 1640 – 1644

uncertainties exist about whether the prevalence for doping is reliable. Socially sensitive issues such as, for instance, illicit drug use, may be under-reported, due especially to the direct approach of these survey tech- niques, as people hesitate to provide compromising information about themselves. Accordingly, which parameters are reliably associated with doping are still a matter of debate. In particular, the assumption that sen- sitive issues such as illicit drug use are associated with doping may be biased by response errors. In order to clarify this point, the present study employs the random- ized response technique (RRT), which requests informa- tion in an indirect manner and therefore minimizes response errors [13,14].

We hypothesized that RRT will confirm the results of our survey with respect to socially non-sensitive issues. In contrast, however, we expected that the RRT would reveal substantially higher prevalences for socially sensitive issues. The outcome of RRT will identify which associa- tions between doping and other parameters could be regarded as reliable. This information will be used to improve the characterization of doping substance users in fitness sports.

METHODS

Sample

A total of 500 fitness centre members were interviewed. These members were recruited from a representative sub- set of 49 fitness centres from the 113 centres that were included in the previous survey [12]. One per cent of the members in each centre were recruited for an interview. Selection of subjects was conducted at different time- points during the opening hours of the different fitness studios. Training directors of the particular facilities were asked to assist with the random choice recruitment of fit- ness centre members, in order to ensure only involve- ment of members of the particular facilities and to maximize the participation response rate. Only one mem- ber refused to participate in this study.

Randomized response technique

The standard version of the unrelated question RRT design was used to acquire information about five sensi- tive issues by presenting sensitive questions in an indirect manner [14]. Each of the five sensitive issues was tested in a separate trial. Thus, the complete interview was com- posed of five trials. In brief, in each trial, one sensitive question and one non-sensitive question unrelated to the sensitive issue was formulated (Table 1). The respondents received a deck of 20 cards which contained 15 replica- tions of the sensitive question and five replications of the innocuous question. In each trial, the respondents were shown the deck so they were fully aware of the two types of questions. They were then asked to turn the deck around, to shuffle it, to draw a card from the shuffled deck and to answer the question on the card drawn with ‘yes’ or ‘no’. This procedure was repeated for the remaining four sensitive questions. Thus, in each of the four remain- ing trials, a new deck with 15 replications of another sensitive question and with five replications of another non-sensitive question was used (see Table 1 for the pair of questions that was employed in each of the five trials).

Because the interviewer did not know whether the respondents had drawn the sensitive or innocuous cards, the respondents could respond honestly without feeling embarrassed. Although the actual question answered by the respondents remained anonymous to the interviewer, the proportion (i.e. dark figure) of ‘yes’ responses with respect to the sensitive question can nevertheless be inferred from the answers of all respondents using the fol- lowing formula:

[

−

)·

where

denotes the proportion of respondents in the sample which replied with ‘yes’ to the drawn card (i.e. irrespective of type of question),

is the probability of drawing the sensitive card from the deck (i.e. in our case

3/4) and

corresponds to the probability of answer- ing the innocuous question with ‘yes’. In our study,

equals 120 of 365.25 for the first three innocuous

Table 1

Results of the personal survey using the randomized response technique.

Sensitive question Unrelated question ‘Yes’ ‘No’ a

95% CI

Have you ever used doping substances?

Is your mother’s birthday in the first 10 days of her birth month?

88 412 0.176 12.5% 8.2–17.4

Have you ever taken illegal drugs? Is your mother’s birthday in the first 10 days of her birth month?

196 304 0.392 41.3% 35.6–47.2

Have you ever used cocaine? Is your mother’s birthday in the first 10 days of her birth month?

96 404 0.192 14.6% 10.1–19.5

Do you smoke? Does your house number end in an even number? 166 334 0.332 27.6% 22.1–33.4 Do you drink alcohol? Does your telephone number end in an even

number? 324 176 0.648 69.7% 63.9–75.3

Addiction,

101

, 1640 – 1644

1642

Perikles Simon

et al.

questions and to one of two for the remaining two (cf. Table 1). The variance of the sampling distribution for

can be estimated from:

Var(

)

[

·(1

−

)]/(

with

denoting the number of respondents in the sample. This variance was used to compute a 95% confidence interval for

. Following the RRT interview, we assessed the follow-

ing parameters, as described previously [12]. In brief, the following parameters were included: (1) age (years), (2) height (cm), (3) weight (kg), (4) sex (m/f), (5) schooling (A-level, which is the highest school education level in Germany, or no A-level), (6) family status (married versus unmarried and divorced), (7) nationality (German or for- eign), (8) training years (years), (9) training frequency (times per week) and (10) sports practised (fitness train- ing, weight-lifting or body-building).

RESULTS

The majority of participants were male (69.4%), the aver- age age was 32.0 (SD

11.0) and the average body mass index was 24.2 kg/m

(SD

3.5). The basic characteris- tics are shown in Table 2.

Table 1 shows RRT results. We estimated the number of unknown cases for the prevalence of doping to be 12.5% (95% CI: 8.2–17.4%). This figure agrees with the

results of our recent questionnaire survey. Table 3 shows a comparison of the outcomes of both studies. Accord- ingly, almost identical results were obtained for the socially non-sensitive issues of nicotine and alcohol con- sumption. However, and most importantly, RRT revealed an approximately threefold higher number of unre- ported cases for the prevalence of socially and legally sensitive issues such as illicit drug use, particularly cocaine use.

The present RRT study confirmed the results of the previous study with regard to the socially non-sensitive issues. Therefore, we used only these non-sensitive parameters to predict doping substance abuse by apply- ing nominal logistic regression analysis to the data from our previous study. A prediction profile for doping sub- stance abuse was calculated (Fig. 1) on the basis of seven non-sensitive predictor variables. An individual is recog- nized by this model as a doping substance user if his or her probability for taking doping substances is above 50%. The sensitivity of the prediction is 40.5, the speci- ficity 96.7 and the area under the receiver operating curve is 0.87.

DISCUSSION

Previous studies used direct interview techniques and questionnaires to reveal the estimated number of unre- ported cases of doping and to investigate the association of doping with other parameters [1,2,4,15]. Although previous studies used comparable methodologies, some studies suggested a strong significant association between doping and illicit drug abuse, while others suggested the opposite or at least no correlation [10,12,16,17]. The present RRT confirmed the estimated number of unreported cases of the socially non-sensitive questions in our previous study but revealed figures about three times higher for illicit drug abuse, particu- larly cocaine abuse. This agrees with a recent Italian study, which found very high levels of cocaine metabo- lites in urban waste water [18]. The amount of metabo- lites found allowed them to conclude that the prevalence of cocaine use within the population must be higher than the numbers reported officially. This finding

Table 2

Basic characteristics of 500 fitness centre visitors.

Variable No. (%)

Sex Female 153 (30.6) Male 347 (69.4)

School education A-levels 231 (46.2)

Family status Married 143 (28.6)

Sport Fitness training 292 (58.4) Weight-lifting 127 (25.4) Body-building 81 (16.2)

Nationality German 419 (83.8)

Training years

6 317 (63.4)

≥

6 183 (36.6)

Training frequency Sporadic 1 (0.2) 1–2 times a week 171 (34.2) 3–4 times a week 261 (52.2) 5–6 times a week 54 (10.8) Daily 13 (2.6)

Table 3

Comparison of survey results using a questionnaire and the randomized response technique (RRT).

Variable Questionnaire RRT

Doping ‘yes’ 13.5% 12.5% Illegal drugs ‘yes’ 15.9% 41.3% Cocaine ‘yes’ 4.7% 14.6% Nicotine ‘yes’ 30.0% 27.5% Alcohol ‘yes’ 77.8% 69.6%

Doping in fitness sports

1643

Addiction,

101

, 1640 – 1644

supports the notion that estimates of prevalence derived from direct interview techniques or voluntary question- naires are too low for sensitive issues such as illicit drug use, particularly the use of cocaine. Accordingly, contradictory results between different studies regarding sensitive issues may be based simply on methodological insufficiencies.

The estimated number of unknown cases for such socially non-sensitive parameters as smoking and alcohol consumption agree with results published previously (Table 3). Surprisingly, the estimated number of unre- ported cases for doping among fitness centre visitors seems to be estimated correctly by direct interview tech- niques. This finding suggests that doping may not be per- ceived as a sensitive issue by fitness centre visitors. One of the potential reasons for this phenomenon may well be a prominent role of the German health care system in pro- viding fitness centre visitors with the doping drugs AAS and Clenbuterol, as well as monitoring their side effects [12].

The present study revealed a 12.5% prevalence of dop- ing among fitness centre visitors, which agrees ade- quately with the figure of 13.5% in our previous study, based on a larger sample from the identical population [12]. Both figures are the highest reported so far, as they include a representative proportion of females who are known to use doping substances to a lesser extent [4,9,15].

RRT is particularly suited to confirm or reject the prev- alence of parameters investigated in previous studies [13,19]. In the case that RRT does not confirm a param- eter of a direct interview study, it is reasonable to assume that any relationships suggested on the basis of this parameter are not reliable. However, parameters con- firmed by RRT may be used to further improve investiga- tion of the associations between these parameters by resorting to the data obtained in conventional interview or question studies. We instanced this procedure by con- structing a model to predict the individual probability for the use of doping substances. In our previous study the

Figure 1

Prediction profiles for taking doping substances. The vertical dotted lines for each of the eight dichotomized variables show its current setting to 0 or 1, with ‘0’

‘no’ for the variables ‘domestic citizen’, ‘A-level’, ‘alcohol’, ‘body building’ or below 23.5 kg/m

for the BMI, below 7 years for ‘training years’ or below four times per week for ‘train- ing frequency’. The horizontal dotted line indicates the current probability for tak- ing doping substances on the

-axis, which is also given as a percentage on the left-hand side for each profile. Profile 1 is the profile where all variables are set (represented by the vertical dotted lines) in such a manner as to obtain the profile with the lowest probability for doping. When a variables setting is changed its legend name appears in bold. The individual lines within the plots for each variable show how the proba- bility for taking doping substances would change when the current value of an individual variable was changed

Profile 1

0.2 %

domestic citizen

A-level alcohol BMI training- years

training- frequency