Abstract
Background. The Aging Males' Symptoms scale (AMS) is an internationally used scale to analyse health related quality of life (HRQoL). The aim of this paper is to provide evidence that the Nigerian AMS scale measures HRQoL with similar accuracy as in other language versions. We also intended to show the severity of complaints of aging in males in advanced old age. More generally, we aim to demonstrate that the Nigerian AMS scale is an applicable, validated, sensitive instrument to measure HRQoL in Nigeria.
Material. We performed a representative survey in Nigeria to get data of the AMS scale in a group of males in old age. The survey was a household-based population sample conducted in March 2006. The Nigerian data were compared with existing data from other European countries. Only community-based data were used for this comparison.
Results and discussion. The internal structure of the AMS (factorial analysis) was sufficiently similar with the comparison group of other countries in Europe to conclude that the scale really measures the same phenomenon. The sub-scores and total score correlations were high (0.8–0.9) but lower among the sub-scales (0.4–0.8). This suggests that the domains are not fully independent. The reliability (consistency) was found to be good and almost identical with European countries.
Mean scores of the Nigerian AMS did systematically differ from data of other European studies. There were much higher because of older age. The same applies for the population reference values obtained in Nigeria. It is the first time that population norm values are available for very old age.
Conclusion. The Nigerian AMS scale is a standardized HRQoL scale that showed good psychometric characteristics (reliability, validity) similar to other international versions. We suggest that the results obtained with the Nigerian AMS scale should be used preliminarily as reference for ‘norm values’ for highest age, i.e. in absence of other data. Confirmation in future studies is needed.
Introduction
The Aging Males' Symptoms (AMS) scale and was originally developed as a symptoms profile scale in Germany to evaluate health-related quality of life scale (HRQoL) Citation[1]. It is a self-administered scale to assess symptoms of aging, i.e. across the full range of ages, to compare the severity of complaints, and to measure therapeutic interventions Citation[1-3].
The AMS scale has a broad international distribution, i.e. currently available in 24 languages around the globe. These versions are available in published form Citation[2],Citation[4],Citation[5] and can be downloaded from the official AMS website (www.aging-males-symptoms-scale.info).
Norm values of the standardized AMS scores (total score and three domain scores), however, were published for few countries until now. In addition, little is published yet about the comparability of the scale across countries, except one recent paper that provided promising impressions, which was based, however, on small numbers for most of the countries Citation[6]. It is important to analyse whether the scale measures similarly in different countries.
The AMS scale was also translated into Yoruba, one of the Nigerian languages, and applied there in a survey of the elderly population. It is the objective of this paper to demonstrate how the AMS scale works in elderly males, where published evidence is lacking. Comparisons will be made with methodological experiences gathered in Europe.
Material and methods
The Nigerian data were gathered in a Nigerian household-based population survey of a sample of the elderly population (age range 60 to over 90 years), speaking Yoruba and ready to participate. The study was conducted in March 2006 (see Appendix 1 for the Yoruba translation). The central aim of the survey was to assess the level of general well-being as well as quality of life among a representative sample of the elderly in the selected areas. The sample consisted of 456 elderly males and 491 elderly women. We analysed for this paper only the participating men. The study was a cross-sectional survey covering three local government areas (LGAs) of Osun State in South-western Nigeria. A multi-stage systematic random sampling method was employed to select the respondents in the LGAs. Interviews were conducted personally face-to-face with selected respondents.
We compared the data of the Nigerian survey in elderly males with existing data from European AMS studies of males aged 40–70 years since data are not published for higher age groups as comparison for our results. The European results were published recently Citation[6] and kindly provided by Heinemann for looking at possible similarities or differences of the Nigerian AMS scale. A similar comparison is available in two recent publications from France Citation[7],Citation[8]. The important difference for the Nigerian database is the absence of very old men in the ‘European database’.
We analysed the most fundamental psychometric characteristics of the Nigerian AMS, i.e., the factorial structure of the AMS, and some aspects of reliability/validity, and compared also the severity of symptoms of aging (population reference values) between the Nigerian sample and the European data. Thereby we used for the analysis of the Nigerian AMS scale identical definitions and cut-off points as were applied in the above mentioned methods published by Heinemann and the two French papers Citation[6],Citation[7],Citation[8].
Important quality requirements for a new scale – here the Nigerian AMS – are to demonstrate that the internal structure of the scale (factorial analysis) in a new language is sufficiently similar with other language versions (such as combined in the European dataset). If the Nigerian scale works similarly compared to the scales used in other countries in Europe, the same domains should be identified and similar factor loadings found. In addition, we should be able to provide evidence that reliability (consistency) is sufficiently good and not different from the same scale in other language regions. We used only the internal consistency measure Cronbach's alpha. In contrast to systematic and random variation, reliability gives an estimate of method-related measurement error that should be low – or the consistency coefficient ‘alpha’ high. We also measured and compared preliminary indicators for validity such as total-domain score correlations because no other scales were used in the survey to estimate criterion-orientated validity.
The analyses were done with SPSS Windows 10.0 and STATA 8.0.
Results and discussion
gives a crude description of age and education of the Nigerian sample of men. In all, 456 elderly males were covered in the study. About a quarter of the respondents were aged 80 years and above. More than 70% of the subjects had no formal educational training and about 43% reside in rural communities.
Table I. Demographic description of the sample of elderly males interviewed in Nigeria.
Internal structure of the Nigerian AMS
We analysed multivariately the internal structure (‘dimensions’) of the Nigerian AMS using the standard technique of factor analysis (main component method with orthogonal rotation) to understand whether the Nigerian scale measured a similar phenomenon as observed in other countries.
shows a reassuring result. The analysis of our data came to a similar structure as was found in Europe. Again, the same three domains were found in the Nigerian scale as were described initially for the German original scale by the statistical analysis. In addition, the proportion of explained variance by the AMS scale is also very similar: 65% and 56% in the Nigerian and European analysis (52% in the original German scale in 1996). The loadings of the 17 items on the 3 factors/domains of the Nigerian AMS are also similar with those from rest of Europe. The factor loadings indicate a good similarity for the psychological, somatic and sexual sub-scale (domain), i.e. they are at least in the same ballpark range. However, besides similarities there are also differences. Three out of the seven items originally belonging to the somatic domain seem to be associated similarly or even more with the psychological factor (marked italic in ). In addition, two of the original five items of the sexual domain were more correlated with the psychological domain in this elderly male sample, i.e. also observed in the European sample. It is not clear if this is a phenomenon of random error or more important. It is not likely that these weaknesses influence the use of the scale in practice since individual items are not used – only average scores of the total scale or the three domains. Moreover, it is unlikely to have negative impact on multinational studies, because intra-individual comparisons over time (before/after treatment) are the main criteria that might not be affected very much. Nevertheless, this could be considered when a revision of the AMS scale is planned sometimes in the future. All the same, the factor analyses provide a confirmation that the internal structure of the Nigerian AMS scale is compatible with the structure described for other countries Citation[6], some less some more convincing.
Table II. Internal structure and domains of the AMS scale. Comparison of the community sample of elderly males of Nigeria and males in Europe Citation[6].
Reliability
shows the internal consistency measured with Cronbach's alpha. The consistency coefficients ‘alphas’ vary around 0.8 in the Nigerian sample. The values are not materially different from the alpha values in the combined European sample, i.e. for the AMS total score as well the three subscales. This is indicative for a very acceptable consistency of the Nigerian AMS scale – and compatible with the comparator Europe.
Table III. Internal consistency coefficients (alpha) for the AMS scale in elderly men in Nigeria compared with Europe: Total score and scores for the psychological, somatic, and sexual sub-scale. Mean and SD of the total score of AMS and of domain scores. Community samples.
However, there are marked differences in the mean values between the aged Nigerian sample and the younger European sample. The mean total score as well as the domain scores were much higher in Nigeria than Europe. It was expected that complaints of aging would be more prevalent in the aged Nigerian sample. A clear increase of the AMS scores with age has been published several times. However, the scale obviously works similarly irrespective of age (see factorial structure and estimates for reliability and validity).
Validity
The first step of validation is usually to multivariately demonstrate the internal structure (‘dimensions’) of a given scale through factor analysis. Reassuring results of the Nigerian scale were reported above.
Another criterion is the total score – domain score correlation. The correlations between subscales (supposed to be independent) should be much lower than the correlations with the total score to which all sub-scales should significantly contribute – the domains are really independent from each other (as suggested by the factorial analysis with orthogonal factors). shows a very similar situation for the Nigerian scale and the European comparator. The domain score correlate highly with the total score (0.8–0.9) in the comparison groups, but there are only somewhat lower correlations among sub-scales (0.4–0.8 in Nigeria and 0.5–0.7 in Europe). This suggests that the sub-scales are not as independent from each other as one would expect them to be according to the applied model. This was also observed in other international studies Citation[6], e.g. in the recent French studies Citation[7],Citation[8].
Table IV. Domain score – Total score correlations of the AMS scale. Comparison between the Nigerian and the pooled European community sample.
Nigerian AMS population reference values
The perceived severity of health-related quality of life in the Nigerian population sample is depicted in . 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. The categories of severity were arbitrarily defined in a large German survey Citation[1] and identically used in other countries.
As one can expect from the much higher mean values for complaints – as discussed above – there is significant shift towards moderate or severe complaints in the Nigerian sample as compared with the European sample (). This table of the Nigerian results can be considered as reference distribution for the AMS scale for males in the uppermost age range, whereas the French results Citation[8] can be used as reference for young men under 40 years of age.
Table V. Norm values of the AMS scale estimated from the Nigerian survey of elderly males (n= 456). Categories of severity of complaints in the total scale and the three domains. European population reference values (n= 5907) of males aged 40–70 as comparison.
Conclusions
The Nigerian AMS scale is a standardized HRQoL scale that showed good psychometric characteristics (reliability, validity) – similar to other international versions. We suggest that the results obtained with the Nigerian AMS scale should be preliminary used as reference for ‘norm values’ for highest age, i.e. in absence of other data. Confirmation in future studies is needed.
Acknowledgements
Thanks to Dr Heinemann for providing a secondary analysis of the European database.
Thanks to Obafemi Awolowo University for providing partial funding for the fieldwork. The contributions of Professors Adekunbi Omideyi, Felicia Oyekanmi and Alfred Adewuyi are appreciated.
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References
- Heinemann L AJ, Zimmermann T, Vermeulen A, Thiel C. A new ‘Aging Male's Symptoms’ (AMS) rating scale. The Aging Male 1999; 2: 105–114
- Heinemann L AJ, Saad F, Thiele K, Wood-Dauphinee S. The Aging Males' Symptoms (AMS) rating scale. Cultural and linguistic validation into English. The Aging Male 2001; 3: 14–22
- Heinemann L AJ, Saad F, Pöllänen P. Measurement of quality of life specific for aging males. Hormone replacement therapy and quality of life, H PG Schneider. Parthenon Publishing Group, London 2002; 63–83
- Heinemann L AJ, Saad F, Zimmermann T, Novak A, Myon E, Badia X, Potthoff P, T'Sjoen G, Pöllänen P, Goncharow N P, et al. The Aging Males' Symptoms (AMS) scale: Update and compilation of international versions. Health and Quality of Life Outcomes 2003; 1: 15
- Convay K, Heinemann L AJ, Giroudet C, Johannes E J, Myon E, Taieb C, Raynaud J P. Harmonized French version of the Aging Males' Symptoms Scale. The Aging Male 2003; 6: 106–109
- Daig I, Heinemann L AJ, Kim S, Leungwattanakij S, Badia X, Myon E, Moore C, Saad F, Potthoff P, Do Minh T. The Aging Males' Symptoms (AMS) scale: Review of its methodological characteristics. Health & Quality of Life Outcomes 2003; 1: 77, Available from: http://www.hqlo.com/content/1/1/77
- Myon E, Martin N, Taieb Ch. The French Aging Males' Symptoms (AMS) scale: A methodological review. Health & Quality of Life Outcomes 2005; 3: 20, Available from: http://www.hqlo.com/content/3/1/20
- Myon E, Martin N, Taieb Ch, Heinemann L AJ. Experiences with the French Aging Males' Symptoms (AMS) scale. The Aging Male 2005; 8: 184–189
Appendix 1
Yoruba translation of the AMS Scale
Bawo ni ose ro pe ilera re se peye to? Gbe ara re si ori iwon (ki aarun je iwon to o ga julo ki ookan si je eyi ti o kere julo)