Experiences with the French aging males' symptoms (AMS) scale

Abstract

Introduction. The aim of this paper is to report on experiences gained from the application of the French version of the Aging Males' Symptoms (AMS) scale, to show correlations with other relevant scales and to provide population reference values for France.

Methods. A representative survey based on an existing, representative population panel was performed in France (963 males, aged 15 + years). Other AMS data from Europe were then used for comparison. Mean scores of the French AMS scale do not systematically differ from the data of other European studies. The population reference values obtained for France strongly suggest that the scale measures can appreciate the level of well-being whatever the patients' age; however, some variables, such as age and family income, can influence the total AMS scores. Comparisons with the generic quality-of-life scale, SF-12, and a depression scale, HAD, have shown good correlations, as have comparisons with the ADAM questionnaire.

Conclusion. The French AMS scale is a standardized, valid and simple health-related quality of life (HRQoL) scale, with results comparable to other published European findings. The results also indicate that the AMS scale can be used to measure and compare HRQoL in those less than 40 years of age.

Keywords:

• AMS
• Aging Males' Symptoms
• health-related quality of life
• questionnaires
• validity

Introduction

Clinical research's increasing interest in aging males prompted recognition of the need to measure health-related quality of life (HRQoL) and associated complaints. The Aging Males' Symptoms (AMS) scale was originally developed and standardized as a symptoms profile scale in Germany, to evaluate the HRQoL scale [1]. It is a self-administered scale, designed to assess symptoms of aging, to compare the severity of complaints and to measure the impact of therapeutic interventions [1],[2],[3].

The AMS scale is internationally well known. It has been translated into 17 languages, following methodological recommendations for linguistic and cultural adaptation. These versions are available in a published form [2],[4],[5] and can be downloaded from the internet (http://www.aging-males-symptoms- scale.info).

Norm values of the standardized AMS scores (total score and three domain scores) have so far only been published for Germany. Furthermore, little has been published yet about the comparability of the scale across countries, except for one recent paper that provided promising impressions [6]. It is important to analyze whether the scale allows similar measures in different countries.

The aim of this paper is to report findings of the French version of the AMS scale compared with other language versions, e.g., to see whether the age distribution of AMS scores is similar when compared to European experience. We also intend to provide reference values for France's AMS scores. Moreover, we will compare results of the AMS scale with a generic quality-of-life scale (SF-12) [7], as has been done in Germany [3], we will correlate it with a depression score (HAD questionnaire) [8] and we will contrast it with a screening tool for androgen deficiency in aging males (the ADAM questionnaire) [9]. This should permit conclusions as to whether the AMS scale is a relevant, valid, sensitive instrument to measure HRQoL, and also illustrate whether there has been a change of symptoms in France, as has been shown for Europe in general [6],[10].

Methodology

Based on a representative national population panel, 963 men aged 15 years and above (age range: 15–90 years old) were recruited for an interview survey (omnibus survey). The survey participants were invited and stratified by the following criteria: sex, age, profession of the head of family, category of town and region (quota method).

The quota method is, first and foremost, an empirical methodology of representative sampling forming. It is based on the following hypothesis: If the sample is structurally similar or identical to the reference field, the results observed by the survey can be extrapolated to this reference population. The method's efficacy in many fields has, of course, been regularly validated by comparing survey data to the reality identified of the general population. The idea is therefore to establish a sample that is structurally comparable to the reference population, based on a certain number of variables for which this structure is known. In general population surveys, the most widely used variables are gender, age, profession, quotas applied in geographical strata (household size and region) and the stratification principle, improving the reliability of estimations. Other variables may be individually used as additional quotas. In practice, the objective is therefore to build up a sample whose final structure, at a marginal level, will be comparable to that of the reference population. As the global sample is made up of the compiling of all individual samples collected by the interviewers, the latter are frequently given an individual quotas plan, particularly in face-to-face methodology, where analysis of the sample structure can only be performed afterwards.

The survey was performed by telephone in two seasons, spring and summer, 2004 (n = 483 in March and n = 480 in July). AMS scores were available for 903 men. Thus, the final response rate for AMS was 93.8 %. In March, all participants were asked to reply to the French version of the AMS scale [5], to the generic quality-of-life questionnaire SF-12 [7] and some demographic questions. Two additional instruments were applied in the second part of the survey in July: a depression scale (HAD questionnaire) [8] and a screening tool for androgen deficiency in aging males (the ADAM questionnaire) [9].

The analyses were performed with explorative statistical methods for purposes of description. Differences among variables were tested with Chi-2-test and the Student t-test. Logistic regression was used to determine the impact of socio-economic status (SES) on the severity of complaints assessed with the AMS scale.

Comparisons between SF-12 scores, HAD depression scores, ADAM and AMS scores were performed using a generalized linear model (GLM) with adjustment for age, sex, family income, region and population density. All statistical analyses were done with commercial statistical packages such as SPSS Windows 10.2, SAS 8.2 and STATA 6.0.

Results and discussion

The survey analysis encompassed 963 men aged 15 and more years. AMS scores were available for 903 men. Although the survey was performed in two blocks, i.e., March 2004 (n = 442) and July 2004 (n = 461), we combined both subsamples because no differences between the AMS total scores or domain scores were observed (data not shown). Both samples could be considered as representative samples of the French population (based on quotas method).

The respondents of the survey were distributed over the five regions of France: Ile-de-France 19%, northwest 23.5%, northeast 22.5%, southwest 11.4% and southeast 23.6% (Table I). The largest proportion of respondents lived in municipal communities, such as Paris and towns with over 100,000 inhabitants (46.2%), followed by those living in rural regions (23.6%). The annual income of the family head (usually the father) was well distributed over the five categories, the highest proportion being in the category of 13,000 to 23,000 Euros, as can be seen in Table I.

Table I.  Demographic description of the population sample of the French survey.

	n	Mean (SD)
Age (years)	963	44.7 (17.3)
	N	Percent (%)
Region
Ile de France	183	19.0
Northwest	226	23.5
Northeast	217	22.5
Southwest	110	11.4
Southeast	227	23.6
Total	963	100
Living area/population density
Rural	227	23.6
Less than 20,000	170	17.6
20,000 – 100,000	121	12.7
More than 100,000	278	28.9
Paris	167	17.3
Total	963	100
Age distribution
Between 15 to 30 years	232	24.1
Between 31 to 40 years	216	22.4
Between 41 to 50 years	169	17.5
Between 51 to 60 years	160	16.6
More than 61 years	186	19.3
Total	963	100
Family income per year (in Euro)
Less than 13,720	189	22.3
13,720 – 23,781	291	34.4
23,782 – 36,586	191	22.6
More than 36,587	175	20.7
Total	846	100

French AMS scores different?

Table II provides details about the average AMS total and domain scores in the French population, and additionally stratified in two age groups. The AMS scale was initially developed for men over 40 years old. If one compares this with combined data from other European countries [6] (France excluded), there seems to be no systematic and clinically meaningful difference in the mean values.

Table II.  Mean AMS scores (SD) for total scale and three domains by age groups. Data from a european database [6] of men aged 40 years for comparison.

	All age groups		Age under 40		Age 40 +		Europe
	n	Mean (SD)	n	Mean (SD)	n	Mean (SD)	n	Mean (SD)
Total score	903	27.7 (9.3)	400	24.7 (7.4)	503	30.1 (10.0)	5907	28.0 (9.4)
Psychological score	903	7.6 (3.0)	400	7.2 (2.6)	503	7.9 (3.2)	5907	7.4 (3.1)
Somatic score	903	12.5 (4.7)	400	11.2 (4.0)	503	13.5 (4.9)	5907	12.6 (4.7)
Sexual score	903	7.6 (3.2)	400	6.2 (2.1)	503	8.6 (3.4)	5907	7.9 (3.4)

If there is no convincing difference in mean values of community-based samples of those over 40 years of age, in Europe and France, the scale obviously measures similar phenomena. This is reassuring for the application of the scale in France. This should also be discussed in the context of analyses that showed a similar factorial structure of the AMS results in France and elsewhere. These methodological results will be separately published soon.

French AMS population reference values

The perceived health-related quality of life in the French population sample is depicted in Table III for all respondents. The degree of impairment is demonstrated by four categories of the AMS total score, and also for the psychological, somatic, and sexual domain, i.e., no, little, moderate and severe impairment. Severe and moderate complaints are less frequent than no or little problems, e.g., for the total score and the domain scores. This table can be considered as reference distribution for the AMS scale in France.

Table III.  Norm values of the AMS scale estimated from the French survey. Classification according to four categories of severity of complaints in the total scale and the three domains. All males aged 15 + years (n = 903) and for international comparisons also men 40 + years (n = 400). European population reference values [6] as a crude comparison can be seen in the right column. (grey = other european data as comparator.)

		France						Europe
		All age groups		Age under 40		40 + years		40 + years
Points	Impairment	n	Percent (%)	n	Percent (%)	n	Percent (%)	Percent (%)
Total scale
−26	No	505	55.9	284	71.0	221	43.9	54.4
27 – 36	Little	243	26.9	82	20.5	161	32.0	28.9
37 – 49	Moderate	122	13.5	28	7.0	94	18.7	12.9
50 +	Severe	33	3.7	6	1.5	27	5.4	3.8
Psychological dimension
−5	No	283	31.3	137	34.3	146	29.0	33.6
6 – 8	Little	368	40.8	166	41.5	202	40.2	42.1
9 – 11	Moderate	104	11.5	19	4.8	85	16.9	14.3
12 +	Severe	148	16.4	78	19.5	70	13.9	9.9
Somatic dimension
−8	No	180	19.9	112	28.0	68	13.5	18.1
9 – 12	Little	360	39.9	185	46.3	175	34.9	39.7
13 – 18	Moderate	260	28.8	78	19.5	182	36.2	30.1
19 +	Severe	103	11.4	25	6.3	78	15.5	12.1
Sexual dimension
−5	No	332	36.8	230	57,5	102	20.3	31.1
6 – 7	Little	235	26.0	100	25,0	135	26.8	24.9
8 – 10	Moderate	178	19.7	44	11,0	134	26.6	24.6
11 +	Severe	158	17.5	26	6,5	132	26.2	19.4

For the higher age group of 40 and older, we can roughly compare the severity pattern in the French population sample with the distribution seen in other European community samples [6]. At face value, there are similar distribution patterns for the severity of complaint in France and other European samples (all respondent were aged 40 and more years), although apparently somewhat lower severity of complaints in countries other than France, if compared superficially.

The large, combined European sample (n = 5907 males) is a relevant comparator to demonstrate that the French AMS scale allows equivalent measures as those versions validated and used in other countries. There is also no evidence of a systematic difference, and different trends were not detected. It is our hypothesis that slight differences might be explained by which, the subjective perception of certain symptoms differs slightly among different populations. Moreover, the possibility that ascertainment bias and other forms of bias might have influenced the distribution cannot be excluded – although we have no evidence for such an allegation.

This, together with a similar factorial structure in France and rest of Europe, is reassuring (data not shown; to be published elsewhere). Nonetheless, it is important that other language versions provide their own reference values for the respective population. This would reinforce the uniformity of the AMS scale as a measure for the same pheno- menon everywhere, and argues in favor of a sufficiently good and validated cross-cultural adaptation of the scale.

The French normal values of the AMS scores are the first reference values published outside Germany since the AMS scale was developed. We think that such population-based information is required from other countries, so as to learn more about intercountry variability and reasons for variation.

Variables influencing the AMS score

We checked some socio-demographic factors such as age, geographical location, population density of locality, family income and body mass index (BMI), to determine whether these have an impact on the AMS score.

We observed a significant correlation between age and the AMS total, somatic and sexual scores but not for the psychological scale in the univariate analysis. This was similar for income. No association was found between geographical location or population density.

These variables, however, might be interrelated, and therefore multivariate analysis required. Persons with higher income might, for example, be older and prefer to live in municipal areas. When all the above-named variables are entered into a multivariate logistic regression model (adjusting for the other variables), age (over 40 years) has a significant association with higher AMS total score: odds ratio = 3.3 (95% confidence interval, 2.5 to 4.9). Higher incomes of the family, however, were associated with significantly lower AMS total score: OR 0.5 (0.4 to 0.7). BMI [OR 1.3 (0.9 to 1.9)] and city type/population density [OR 0.8 (0.6 to 1.2)] were not significantly associated with the AMS total score.

Thus, comparisons of AMS scores between different groups should be adjusted for age and family income, at least. One can hypothesize that other marker for socio-economic status (such as education) might be as useful as income for adjustment.

Associations with other scales

Correlation with generic quality of life (SF-12)

The SF-12, together with the AMS scale, was administered to all participants surveyed (in March and July). The correlation coefficients range between 0.3 and 0.5 (inverse correlation). The greater the severity of symptoms measured with AMS, the lower the global health status is; the mean values of the physical dimension of the SF-12 (see mean values in Table IV). The same applies for the mental dimension of the SF-12.

Table IV.  Comparisons of the AMS total score with: Quality of life (measured by SF-12); depression (measured by HAD); suspicion of androgen deficiency in aging males (measured by ADAM screener [9] in men over 40 years). Upper part: Mean values (SD) of scores of other scales (only applied in the second part of the survey) stratified by categories of severity of complaints measured by the AMS scale. Lower part: Mean values (SD) of the AMS total score and 3 AMS domains in screening positive and screening negative men according to ADAM scale.

		AMS scale: Severity of symptoms
		No/little	Mild	Moderate	Severe	P-value*
SF-12
PCS	n	497	236	121	32
	Mean (SD)	53.3 (4.9)	49.1 (8.2)	45.5 (9.6)	40.1 (9.7)	< 0.0001
MCS	n	497	236	121	32
	Mean (SD)	51.5 (6.7)	47.1 (8.4)	40.9 (10.3)	35.5 (10.1)	< 0.0001
HAD	n	266	116	53	18
depression score	Mean (SD)	2.4 (2.2)	3.8 (2.5)	6.0 (2.8)	11.1 (4.8)

		Screening scale for androgen deficiency (ADAM)^§
AMS		Screening positive	Screening negative	P-value*
N		167	88
Total scale	Mean (SD)	33.4 (10.3)	22.9 (4.9)	< 0.0001
Psychological domain	Mean (SD)	8.4 (3.4)	6.3 (1.9)	< 0.0001
Somatic domain	Mean (SD)	15.1 (5.1)	10.4 (3.1)	< 0.0001
Sexual domain	Mean (SD)	9.9 (3.5)	6.2 (1.5)	< 0.0001

*All p-values adjusted for age, family income, region, and population density (GLM procedure).

The physical domain of the SF-12 was highest correlated with the somatic domain of the AMS (r = 0.49) but very similar to the total score (r = 0.47). The mental domain of the SF-12 was particularly associated with the psychological subscale of the AMS (r = 0.54), and almost identical to the total AMS score (r = 0.51). In a previous study [3],[4], a similar correlation of the AMS scale with the SF-36 was reported (r = 0.49). Thus, the association with generic QoL in France seems to be similar to another European country, i.e., pointing again in the direction of compatibility of the AMS across countries.

Association with a depression scale (HAD)

The depression section of the HAD scale was used in the second part of the survey only. We found significant correlations between the depression score of the HAD scale and the AMS total score (r = 0.62; p < 0.0001), the somatic domain (r = 0.58; p < 0.0001), the psychological subscale (r = 0.55; p < 0.0001) and the sexual domain (r = 0.45; p < 0.0001). The mean depression score of the HAD significantly increased with increasing severity of complaints determined by the AMS scale (data not shown). This finding was expected, since higher severity of complaints should have an impact on perceived depression and vice versa.

Association with the results of a screening tool for androgen deficiency (ADAM)

The ADAM scale of Morley et al. [9] was only applied in the second part of the survey, and only for those males over the age of 40 (see Table IV for analyses of results).

Although this screening tool is – like the AMS – based on a profile of symptoms, the categories ‘screening positive’ and ‘screening negative’ should be associated with the severity of symptoms measured with the AMS scale. Sixty-five per cent of the respondents were labelled screening positive and 35% were not positive, according to ADAM. ADAM positive men reported significantly higher mean AMS scores for the total scale, the psychological, somatic, and sexual domains. This finding would support the hypothesis that the ADAM questionnaire and the AMS scale measure a similar phenomenon, although the objectives of both instruments differ.

The associations of the French AMS results, in conjunction with findings of other relevant scales, point to the direction that the AMS scale measures health-related quality of life, is linked to depression and does not contradict results of a screening scale for androgen deficiency, the ADAM questionnaire.

Conclusion

The experience with the French AMS scale is promising: the mean scores of the French AMS did not systematically differ from data obtained in other European centers. Population reference values are rarely reported, except for Germany and now for France. They seem to be grossly compatible. Moreover, reference values for the age group under 40 years were, for the first time, reported in this French study. The data suggest strongly that the AMS scale measures a similar phenomenon across age. Impact from other variables on the AMS score was also demonstrated: the score increases significantly with age, and also independently with declining family income.

Altogether, the findings suggest that the French AMS scale is a standardized, valid, and simple HRQoL scale with results comparable to experiences published in other European countries. The reference values of the AMS for the French population demonstrated similarly with European experience. We can suggest that the AMS scale should also be used in age groups under 40 years, to measure and compare health-related quality of life in males.

Acknowledgements

We thank Ipsos Santé Poll Institute for performing the fieldwork for the French survey over two periods.