What Works for Patients in Outpatient Treatment for Alcohol Addiction? An Explorative Study into Clients’ Evaluation of Subjective Factors and Therapy Satisfaction

Abstract

This explorative survey investigated clients’ evaluation of therapy elements and other supportive factors within a randomized controlled trial. The treatment of patients with alcohol dependence consisted of pharmacotherapy (acamprosate/naltrexone/placebo) and biweekly medical management (MM). Forty-nine study participants were surveyed with a questionnaire to measure both the patients’ satisfaction with the therapy and the subjective assessment of treatment elements and supportive factors.

Study participants were highly satisfied with the treatment. The supportive factors previously identified by Orford et al¹ were confirmed. ‘Pharmacotherapy’ was rated significantly less effective than ‘MM’ and ‘global study attendance’ (P < 0.001). The significant differences in the evaluation of treatment elements point to a preference for regular low-key contacts rather than for medication. Such contacts based on MM could be a useful intervention in clinical care, and its effectivity should be examined more closely in further research.

Keywords

alcohol addiction outpatient treatment therapy satisfaction supportive factors subjective assessment

Introduction

Numerous studies inquiring into effective treatment for alcoholism exist. Currently 865 studies on the treatment of alcohol use disorders are listed worldwide on clinicaltrials.gov. Most of these inquiries are attempting to investigate which specific factors, (e.g., therapy modalities, different medication) are more important or helpful. However, what do patients consider as helpful and important? Does the answer to this question have any relevance for the effectiveness of treatments for alcohol use disorder (AUD)?

Findings from neuroimaging studies strongly suggest that subjective mental processes, such as beliefs and expectations, significantly influence human behavior.² Kocsis et al³ for example, found that patients’ preferences moderated treatment response in patients with depression. In another study, AUD patients with positive expectations for their future showed a more successful course of treatment.⁴ Yet presently it is not common practice to incorporate patient perspectives or preferences into AUD treatment outcome.^1,5

Orford et al^1,6,7 conducted a qualitative investigation of clients’ perspective on change during treatment for AUD. The analyses of the interviews led to the formulation of a model of change including factors that the participants considered important, e.g., family and friends’ support, acting differently, thinking differently, etc.

Additionally, there is one common factor whose importance for treatment success is virtually undisputed amongst specialists: the therapeutic alliance.^8–11 Therapeutic alliance increases participants continuation in therapy, compliance with therapy procedures, and outcome. A meta-analysis on 79 studies showed a moderate correlation of therapeutic alliance and outcome.¹² The relationship between patient and therapist influences compliance to a great extent.¹³ In the treatment of alcohol dependence, the clients’ perception of the therapeutic alliance and satisfaction with Medical Management treatment is predictive for drinking outcomes.¹⁴

A particularly contentious area concerns the relative merits of pharmacotherapy and psychotherapy in the treatment of psychiatric disorders.¹⁵ In depression research psychotherapy and pharmacotherapy did not show strong differences in effect sizes, suggesting that treatment choice should be based on criteria such as contraindications, treatment access, or patient preferences.¹⁶

Surprising results from the large American study Project COMBINE, where psychotherapy was shown to be less effective than Medical Management with placebo,¹⁷ could well be due to the participants’ attitudes towards treatment or what they perceived to be helpful. Participants might have been disappointed not to receive medication, thereby having a negative expectation. Taking pills could potentially function as a positive reinforcement and increase the motivation to stay abstinent.

The present survey wanted to investigate patients’ subjective evaluation of treatment quality and therapy satisfaction as well as supportive factors in a multi-centre, double-blind, randomized, stepped-care therapy project for patients with an alcohol dependence. Of specific interest was which therapy elements were perceived as particularly effective and if patients identified any new factors.

Methods

Participants

This explorative survey was conducted with a sequential sample of study participants attending the first follow-up visit three months after the end of the treatment phase within Project PREDICT at the participating study centers in Mannheim, Freiburg, Tuebingen and Regensburg from October 2006 to August 2007. The study was approved by the Institutional Review Boards of the participating study centers and all participants provided written informed consent.

Project PREDICT

Project PREDICT was a 6-year randomized double-blind and placebo controlled clinical trial in seven German study centers and consisted of two consecutive RCTs in a stepped care model. A total of 427 participants were recruited from inpatient detoxification programs at the respective centers. The treatment phase of step one lasted six months and consisted of biweekly visits following a manualized protocol known as Medical Management (MM) plus pharmacotherapy (acamprosate or naltrexone or placebo) in the first three months.^14,18 Follow-up visits were scheduled every three months for the duration of a year.

If a patient relapsed within the first year of step one a new 4-month treatment phase was offered in the second step.¹⁹ It consisted of MM plus the same medication as in step one, with one randomly selected group receiving an added manualized cognitive-behavioral intervention, Alcoholism Specific Psychotherapy (ASP).^20,21 A total of 109 participants were recruited for this step. For an in depth presentation of the rationale and design of Project PREDICT see Mann et al.¹⁸

Instruments

Data collection was conducted by means of a semi-structured questionnaire, comprising questions developed by the authors and the Helping Alliance Questionnaire (HAQ) in the German translation by Bassler et al.^22,23 The latter includes twelve items and measures two subscales ‘satisfaction with therapeutic relationship’ and ‘satisfaction with therapeutic outcome’ as well as the global therapeutic outcome. It further contains two questions concerning areas of improvement or deterioration in a free text format.

The part of the questionnaire designed by the authors was developed through a group review process and included three areas of inquiry: subjective assessment of treatment elements (SATE), supportive factors (SUFA), and a free text field as an option for general feedback. SATE factors were derived from the treatment elements MM, study medication and ASP, and were augmented with the global factor ‘study participation’ as well as the interventions ‘self-help group’ and ‘other unspecified treatments/therapies’ (see Table 1). Since it is conceivable that treatment elements could be experienced as hindering or even damaging, the evaluations were to be recorded on a bipolar rating scale. The seven-point Likert-scale was anchored with verbal markers from ‘very debilitating/ harmful’ to ‘very effective/supporting’.

Table 1

Questions in the questionnaire.

Questions introducing rating-scale evaluation
HAQ	Please evaluate the treatment outcome and relation to your therapist according to the following statements. Please mark the field that fits best to what extent you agree or disagree with the statement. Just answer spontaneously, as there are no ‘right’ or ‘wrong’ responses. (Below statements with rating scales) Global therapy outcome compared to start of treatment: (Below rating scale)
SATE	The treatment you received during the last six months consisted of regular conversational contacts every two weeks as well as pharmacotherapy with the study medication. how do you evaluate the particular treatment elements? Please mark the field that fits best how debilitating/ harmful or effective/supporting you found the particular treatment element. (Below treatment elements with rating scales)
SUFA	Perhaps there are other supportive factors influencing your treatment outcome. How do you evaluate the following supportive factors? Please mark the field that fits best to what extent you agree or disagree with the statement. (Below statements on supportive factors according to the Orford-study)
Free text field questions
HAQ	‘I feel that i have improved a lot in the following areas’: ‘I feel that i am doing badly in the following areas’:
SUFA	Were there any other supportive factors not yet listed? if yes, which ones? Please describe any factor in a separate text field.
Final comments	Is there anything else that you would like to comment on?

The Supportive Factors (SUFA) were based on the categories derived from results of the UKATT study by Orford et al.¹ The categories ‘family/friend support’, ‘thinking differently’, ‘acting differently’, ‘seeing the benefits’, ‘catalyst’ and ‘down to me’ were presented in the form of statements which were to be evaluated on a five-point rating-scale of agreement. A free text field was added, so that further relevant factors could be identified.

The severity of addiction was measured at study inclusion with the German version of the Structural Clinical Interview for DSM III-R (SCID) and the Alcohol Use Disorders Identification Test (AUDIT).²⁴

Analysis

In a first analysis the mean, median, and standard deviation of the ratings for the therapy satisfaction of the HAQ scales and the various SUFA and SATE factors were calculated. Furthermore a dropout analysis was conducted to see whether the patient characteristics of patients participating in the survey were different from those who had previously discontinued Project PREDICT.

Non-parametric tests were used based on the assumption that most variables did not have a normal distribution: the Friedman test as analysis of variance and the Wilcoxon rank sum test. A logistic regression analysis for the verification of SUFA and SATE factors as predictors for therapy satisfaction was renounced due to the extremely skewed distribution of the target variable. The number of patients reporting rather low satisfaction was too small to warrant statistical testing.

The qualitative data collected in the free text fields were categorized and coded by two raters in several steps, whereby the categories were derived inductively from the material. In contrast to the usually extensive text material in qualitative designs, only concise textual statements were generated. The analytic summary of contents was based on Mayring's²⁵ contents analysis, which presupposes a summary of the basic material by means of paraphrasing, generalization, and reduction.

Results

Forty-nine of 51 study participants attending a follow-up visit three months after the 6-month treatment phase completed the survey questionnaires (see Fig. 1). Sample characteristics are listed in Table 2. AUDIT scores of all subjects were ≥8 (cut-off indicating hazardous and harmful alcohol use), and 83% ranged >20 (cut-off which warrants further diagnostic evaluation for alcohol dependency).²⁶ Fifteen patients visited a self-help group, four patients took part in the alcoholism-specific psychotherapy (ASP), and ten patients received some other unspecified external therapy. Free text statements were made in 27 (55%) of the 49 completed questionnaires.

Table 2

Baseline characteristics of the study population.

Patient characteristics (n = 49)
Demographic characteristics
Sex; N
Male	32
Female	17
Age; years
Mean (SD)	44.27 (7.9)
Problem severity
AUDIT
Mean (SD)	25.9 (6.9)
SCID^*
Mean (SD)	2.47 (0.5)

Notes:

Score range from 1 to 3 (1 = slight; 2 = moderate; 3 = severe). Abbreviation: SD, standard deviation.

Figure 1

Context of the present survey within project PREDICT and numerical breakdown of the sample.

The dropout analysis showed no significant differences between patients attending the first follow-up visit and individuals discontinuing Project PREDICT, except gender (17 women in the responder group vs. five women in the discontinuation group, P = 0.015).

Therapy satisfaction

Therapy satisfaction according to the HAQ showed the following results: the ratings of the subscale ‘satisfaction with therapeutic outcome’ on a scale of 1 to 6 resulted in a mean of 5.06 (SD = 0.852) and a median of 5.10 (n = 48). The subscale ‘satisfaction with therapeutic relationship’ showed a mean of 5.31 (SD = 0.841) and a median of 5.33 (n = 49). Figure 2 shows the rating of ‘global therapy outcome’. A large majority (44 out of 49) described the global therapy outcome compared to start of treatment as clearly or greatly improved (mean = 2.37; SD = 0.782).

Figure 2

Frequency distribution of rating of global therapy outcome according to the helping Alliance Questionnaire (HAQ).

Free Text comments

In the free text field of the HAQ inquiring into improvement, responses were entered by 26 out of 49 persons (53%). Eight statements were coded as ‘self-concept’ (e.g., “self-assurance”, “self-esteem”, “love for myself”). Seven statements were coded as ‘global improvement’ (e.g., “everything”, “altogether happy”). Seven statements were classed as ‘social relationship’ such as partnership, family and friends relations. ‘Body and health’ (e.g., “healthier living”, “bodily and mental”) were mentioned six times, and five participants stated improvements in ‘social skills’ (e.g., “am able to distance myself better from others”, “can stand up for what I want”). Four statements were coded as ‘behavior/skills‘ (e.g., “approved strategies for staying abstinent”) and two as ‘insight’ (e.g., “believe I understand how it could have gone so far”) and ‘quality of life’. Further statements were categorized as ‘job’, ‘motivation’ and ‘prospects’.

Two persons wrote responses to the HAQ question asking for areas of deterioration. These were coded as ‘relapse’ and ‘process of change’ (e.g., “it is often difficult not to follow old habits and to constantly control oneself, but I am not worse off because of this”).

Six participants entered comments in the general feedback free text field. Those were categorized as ‘therapy feedback’ (e.g., “I could probably not have made it without therapeutic assistance”, “I learned a lot in the therapy”, “this institution gives me strength!”), ‘therapy concept advice’ (“I would recommend including biweekly appointments in the therapy program after discharge, so you know where to turn”), and ‘burdening factors’ (i.e., unemployment).

Subjective Assessment of Treatment Elements(Sate)

Table 3 shows the results of the descriptive and inferential statistics for the subjective assessment of treatment elements (SATE). We observed statistically significant differences between ‘pharmacotherapy’ and ‘Medical Management’ as well as ‘pharmacotherapy’ and ‘global study attendance’ using the Wilcoxon rank sum test (P < 0.001). Since only four of the patients participating in the survey were randomized to receive psychotherapy in step two of Project Predict, an evaluation of the treatment elements of these participants could not be included.

Table 3

Subjective assessment of treatment elements (SATE) on a scale from -3 (‘very debilitating/harmful’) to +3 (‘very effective/supporting’).

	Medical management	Pharmacotherapy	Global study attendance
Mean	2.64	1.15	2.49
SD	0.568	1.383	0.718
Mean rank	2.44	1.36	2.20
Wilcoxon rank sum test
	Pharmacotherapy – Medical management	Global study attendance – Medical management	Global study attendance – Pharmacotherapy
Z	-5.072	-1.807	-4.681
Asymptotic significance (two-sided)	<0.001	0.071	<0.001

Abbreviation: SD, standard deviation.

Supportive Factors(Sufa)

Table 4 presents the results for the Supportive Factors (SUFA). A Wilcoxon rank sum test showed that ‘catalyst’ differed from all other items (P < 0.001): ‘catalyst’ and ‘seeing the benefits’ (Z = -4.942), ‘catalyst’ and ‘thinking differently’ (Z = -4.911), ‘catalyst’ and ‘acting differently’ (Z = -4.666), ‘catalyst’ and ‘family/friend support’ (Z = -4.599), ‘catalyst’ and ‘down to me’ (Z = -4.044).

Table 4

Rating of supportive factors according to the UKATT-findings by Orford et al¹ on a scale of agreement from 0 (‘not at all’) to 4 (‘absolutely’).

	Supportive factors (SUFA)	Mean (mean rank)	SD
1	Seeing the benefits	3.36 (4.00)	0.870
2	Thinking differently	3.34 (3.90)	0.731
3	Acting differently	3.23 (3.71)	0.960
4	Family/friend support	3.17 (3.85)	1.167
5	Down to me	3.04 (3.46)	0.955
6	Catalyst^***	1.66 (2.08)	1.672

Notes:

***

Wilcoxon rank sum test: difference of ‘catalyst’ to all other items, P < 0.001.

Abbreviation: SD, standard deviation.

Nine subjects entered comments in the free text field asking for any other supportive factors not yet listed. ‘Social support’ (e.g., “support through employer”, “conversation with other patients”) was mentioned four times and ‘job’ (e.g., “new job”, “I was able to keep my job”) three times. ‘Body and health’ (e.g., “pregnancy/birth”) and ‘self-concept’ (e.g., “self-assurance”) were each mentioned twice, and ‘spirituality’ once.

Discussion

The principal aim of this survey was to assess the clients’ subjective evaluation of the treatment elements and other supportive factors in Project PREDICT utilizing the HAQ (for therapeutic alliance), the SUFA (for supportive factors), and the SATE (for the subjective assessment of treatment elements). The HAQ showed a high satisfaction with the therapeutic relationship and therapeutic outcome. The SUFA confirmed the supportive factors found by Orford et al in the UKATT study.¹ The SATE indicated that patients rated regular contacts in the form of MM and global study attendance as significantly more effective for the success of treatment than medication. These findings about the clients’ perceptions match a host of other publications emphasizing the importance of common factors over specific factors.^6,7,9,10

The HAQ results show that patients who completed treatment and continued in follow-up were highly satisfied with the therapy. The participants regarded the MM interactions as an important positive experience and source of treatment satisfaction. None of the participants chose a negative rating on the bipolar scale for ‘global therapy outcome’. Furthermore, participants evaluated the therapeutic relationship within the MM contacts as highly positive. This result fits well with the large body of literature stressing the importance of the therapeutic bond as a major factor in treatment outcome.

The supportive factors like ‘thinking differently’, ‘family and friends support’, ‘acting differently’, ‘down to me’, ‘seeing the benefits’ as identified by Orford et al¹ received high approval ratings, except for the factor ‘catalyst’. The low approval for ‘catalyst’ could be due to the German translation of ‘catalyst’ as ‘Schluesselerlebnis’, which implies a singular event in the sense of ‘crucial experience’. The picture changes when categories based on free text statements are attributed to the category ‘catalyst’. Categories like ‘job’ (“new job”, “keeping my job”, “job endangered”) as well as “my health” and “pregnancy/birth” are pointing to ongoing or prolonged circumstances as motivating triggers for a willingness to change and to undergo therapy. The factors mentioned in the SUFA category ‘social support’ may be attributed to the UKATT factor ‘family/friend support’, if this factor was defined more generally to also include the support of employers and fellow patients. No other new factors were identified.

The results of the subjective assessment of the different treatment elements (SATE) indicate that the participants valued the regular low-key interactions provided by MM as an important and effective intervention. One of the most prominent results was that the 13 MM contacts within the treatment period were regarded by the participants as significantly more effective than the pharmacological treatment. Therefore one could reasonably assume that for certain groups of patients a targeted offer of such low-key contacts is a useful intervention in clinical care, and its efficacy should be examined more closely in further studies. It would seem particularly useful to investigate its possible efficacy in its own right, since MM or clinical management are frequent components of clinical studies.

It should be pointed out that the explorative nature of the survey strongly limits the possibility to generalize the findings. Two additional limitations should be addressed:

Even though we had a very high return rate (49 of a possible 51 participants completed the survey) only patients actively participating provided information. Therefore we cannot make any inferences about the opinions of those individuals who discontinued before the 3-month follow up visit.

Concerning the study instruments it should be stated that, aside the use of the HAQ, we constructed only face valid questionnaires to evaluate the subjective assessment of treatment elements (SATE) and supportive factors (SUFA). Due to resource restrictions we were not able to establish validity and reliability by the customary validation procedures through standard samples.

In summary, our paper supports other reported findings concerning the presence and importance of common factors in the treatment of AUD. Furthermore, for clinical practice it seems to suggest that even a medication-only therapy of alcohol dependence should always include regular low-key contacts such as MM. The independent efficacy of such low-level contacts based on MM or clinical management should be tested in future research.

Disclosures

This manuscript has been read and approved by all authors. This paper is unique and not under consideration by any other publication and has not been published elsewhere. The authors and peer reviewers report no conflicts of interest. The authors confirm that they have permission to reproduce any copyrighted material.

Footnotes

Acknowledgements

We thank the members of staff in participating study centers for supporting this investigation: Central Institute of Mental Health Mannheim, Clinic of Addictive Behavior; University Hospital of Tuebingen, Dept. of Psychiatry and Psychotherapy; University Medical Center Regensburg, Dept. of Psychiatry. This study is part of Project PREDICT which was conducted within the framework of Baden-Wuerttemberg Addiction Research Consortium and supported by grant no. FKZ 01EB0112 by the German Federal Ministry of Education and Research. Parts of the results were presented at the annual project report for the Baden-Wuerttemberg Addiction Research Consortium, and in a poster abstract at the 1. German Addiction Research Congress, June 2008.

References

Orford

Hodgson

Copello

The clients’ perspective on change during treatment for an alcohol problem: qualitative analysis of follow-up interviews in the UK Alcohol Treatment Trial.

Addiction. 2006; 101: 60–8.

Beauregard

Effect of mind on brain activity: evidence from neuroimaging studies of psychotherapy and placebo effect.

Nord J Psychiatry 2009; 63: 5–16.

Kocsis

J.H.

Leon

A.C.

Markowitz

M.D.

Patient preference as a moderator of outcome for chronic forms of major depressive disorder treated with nefazodone, cognitive behavioral analyis system of psychotherapy, or their combination.

J Clin Psychiat. 2009; 70(3): 354–61.

Schuhler

Jahrreiss

Wagner

Intervention bei Alkohol- und Medika-mentenmissbrauch: Behandlungserfolg und Gruppeninteraktionsprozesse. In: Bassler

, editor. Wirkfaktoren von Stationärer Psychotherapie/Mainzer Werkstatt über Empirische Forschung von Stationärer Psychotherapie 1998. Gießen: Psychosozial-Verlag; 2000.

Howard

Thornicroft

Patient preference randomised controlled trials in mental health research.

Brit J Psychiat. 2006; 188: 303–4.

Orford

Hodgson

Copello

Krishnan

deMadariaga

Coulton

What was useful about that session? Clients’ and therapists’ comments after sessions in the UK Alcohol Treatment Trial (UKATT).

Alcohol Alcoholism. 2009; 44: 306–13.

Orford

Hodgson

Copello

Wilton

Slegg

To what factors do clients attribute change? Content analysis of follow-up interviews with clients of the UK Alcohol Treatment Trial.

J Subst Abuse Treat. 2009; 36(1): 49–58.

Horvath

A.O.

Luborsky

The role of the therapeutic alliance in psychotherapy.

J Consult Clin Psych. 1993; 61: 561–73.

Lambert

M.J.

Early Response in Psychotherapy: further evidence for the importance of common factors rather than “placebo effects”.

J Clin Psychol. 2005; 61: 855–69.

10.

Elkins

D.N.

Empirically supported treatments: the deconstruction of a myth.

J Humanist Psychol. 2007; 47: 474–500.

11.

Priebe

McCabe

Therapeutic relationships in psychiatry: the basis of therapy or therapy in itself?

Int Rev Psychiatr. 2008; 20: 521–6.

12.

Martin

D.J.

Garske

J.P.

Davis

M.K.

Relation of the therapeutic alliance to outcome and other variables: A metaanalytic review.

J Consult Clin Psych. 2000; 68: 438–50.

13.

Koechel

Fischer

U.C.

Krieger

Tigiser

Faath

Evaluation des Therapieprozesses und der Therapieergebnisse der Psychosomatischen Rehabilitationsbehandlung in der Eifelklinik Manderscheid. In: Bassler

, editor. Wirkfaktoren von Stationärer Psychotherapie/Mainzer Werkstatt über Empirische Forschung von Stationärer Psychotherapie 1998. Gießen: Psychosozial-Verlag; 2000.

14.

Ernst

D.B.

Pettinat

H.M.

Weiss

R.D.

Donovan

D.M.

Longabaugh

An intervention for treating alcohol dependence: relating elements of medical management to patient outcomes with implications for primary care.

Ann Fam Med. 2008; 6: 435–40.

15.

Klein

D.F.

Flawed meta-analyses comparing psychotherapy with pharmacotherapy.

Am J Psychiat. 2000; 157: 1204–11.

16.

Pinquart

Duberstein

P.R.

Lyness

J.M.

Treatments for later-life depressive conditions: a meta-analytic comparison of pharmacotherapy and psychotherapy.

Am J Psychiat. 2006; 163: 1493–501.

17.

Anton

R.F.

O'Malley

S.S.

Ciraulo

D.A.

Combined pharmacotherapies and behavioral interventions for alcohol dependence: the COMBINE study: a randomized controlled trial.

JAMA. 2006; 295: 2003–17.

18.

Mann

Kiefer

Smolka

Searching for responders to Acamprosate and Naltrexone in alcoholism treatment rationale and design of the predict study.

Alcohol Clin Exp Res. 2009; 33(4): 674–83.

19.

Berner

Guenzler

Frick

Finding the ideal place for a psychotherapeutic intervention in a stepped care approach—a brief overview of the literature and preliminary results from the Project PREDICT.

Int J Methods Psychiatric Res. 2008; 17(S1): S60–4.

20.

Brueck

Mann

Alkoholismusspezifische Psychotherapie—Manual mit Behandlungsmodulen. Köln: Deutscher Ärzte-Verlag; 2007.

21.

Brueck

Frick

Loessl

Psychometric properties of the German version of the Motivational Interviewing Treatment Integrity Code (MITI-d).

J Subst Abuse Treat. 2009; 36(1): 44–8.

22.

Luborsky

Principles of Psychoanalytic Psychotherapy. A Manual for Supportive-Expressive Psychotherapy. New York: Basic Books; 1984.

23.

Bassler

Potratz

Krauthauser

Der “Helping Alliance Questionnaire” (HAQ) von Luborsky.

Psychotherapeut. 1995; 40: 23–32.

24.

Saunders

J.B.

Aasland

O.G.

Babor

T.F.

de la Fuente

J.R.

Grant

Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption—II.

Addiction. 1993; 88: 791–804.

25.

Mayring

Qualitative Inhaltsanalyse. Grundformen und Techniken. 6th ed. Weinheim: Beltz; 2006.

26.

Babor

T.F.

Higgins-Biddle

J.C.

Saunders

J.B.

Monteiro

M.G.

AUDIT: The Alcohol Use Disorders Identification Test. Guidelines for Use in Primary Care. Geneva: World Health Organization; 2001.