Descripción de los TCA
1. ANOREXIA NERVIOSA: DESCRIPCIÓN PARA LOS EDUCADORES
Our findings on socio-demographic factors are consistent with data from the national census (43, 44) and the associations between summary measures of the NCD risk factors appear sociologically and biologically plausible. These findings suggest that the national estimates had some evidence of construct and associative validity, and that the STEPS methodology (5)
is generally acceptable for use in Vietnam but with caution needed in the interpretation of the associations of summary estimates due to characterization of tobacco smoking and
hypertension as per STEPS protocols. For tobacco use among Vietnamese men, current smoking was negatively associated with raised blood pressure and glucose because those at highest risk were ex-smokers. Our group identified the hypertension phenomenon previously in a survey in Can Tho (45), and proposed that this was likely due to smokers being prompted to quit by a diagnosis of hypertension. The STEPS core instrument refers exclusively to current smokers and the STEPS report template requires reporting only of the proportion of current daily smokers and their years of smoking and quantities smoked, however. These findings should encourage those using the STEPS instrument to collect, and report, information on ex-smokers to more accurately represent the behavioural profile of their populations. For hypertension, our findings demonstrate that raised blood pressure defined as SBP ≥140 mmHg and/or DBP ≥90 mmHg may lead to implausible provincial-level
correlations with summary measures of other risk factors including PA, mean BMI and cholesterol. This definition does not account for blood pressure treated by antihypertensive medication or other means. The prevalence of uncontrolled raised blood pressure is an important health system indicator, and the findings show that the uncontrolled high blood pressure definition of hypertension results in unlikely associations with other risk factors. This problem was resolved by including those using medication for, or previously diagnosed with, hypertension in the definition of raised blood pressure. These results suggest that those using STEPS protocols need to consider the definition of raised blood pressure that is appropriate for their population, and to be aware that the use of the recommended definition may cause spurious associations. Furthermore, although the validity of the summary estimates for inter-country comparisons was tested through the prism of inter-province comparisons within one country, and to the extent that the results can be generalised in this way, the findings hint that the STEPS methodology is appropriate for the intended purpose of cross- cultural comparisons.
Our estimates of quantities of alcohol consumed expressed in terms of standard drinks had some evidence of construct validity in terms of associations with education and tobacco smoking that were consistent with previous findings of studies in Asian populations (46, 47). The findings accord with the evidence that there is a strong link between the number of standard drinks consumed and the mean levels of BP and the prevalence of hypertension in both Asian (48-51) and Western populations (50, 52-54). The evidence of construct validity
Chapter 7. Summary, implications and future research 154 154 was indirect and limited, however. It was based on the observation that prediction of those outcomes by variables representing information on quantities measured by standard drinks was not worse than prediction by variables representing information on frequency of consumption. Most of the improvement in model calibration and subject discrimination in mathematical models of the relationship between blood pressure/hypertension and alcohol consumption was provided by binary responses to questions on whether or not alcohol had been consumed during a specific period. Because the purpose of the STEPS methodology is to obtain small amounts of useful information on a regular and continuous basis when resources are limited, our findings suggest that asking further information on the quantities consumed in terms of standard drinks imposed an additional burden on subjects that is not justified.
That one-in-six respondents reported unrealistically high values of PA in response to GPAQ, and an additional 2% provided incorrect information, shows that nearly 20% of Vietnamese respondents had difficulty coping with the questionnaire. Those who made such reporting errors were men and particularly the less well-educated among them, younger persons, those who reported a second type of work activity and residents of rural areas who did not, those of non-Kinh ethnicity, and persons from low-income households. Over-estimation of self- reported PA in response to the GPAQ questionnaire when administered in the Vietnamese population has been identified previously (55). Two possible sources of this are the Western concepts of intensity and continuity of effort, and the reference period characterised as a typical week. Pointing to the first of these explanations is the finding that around 10% of respondents reported levels of vigorous activity that improbably required energy expenditure every day of a typical week of the past year in excess of the average energy intake (2100 kcals) of Vietnamese people (56, 57). Either these respondents misunderstood what is required in moderate and vigorous activity, or they were unduly influenced in their reporting by recent bouts of high activity. The second explanation is in accordance with a previous finding by our group (58) that the stability of work patterns influences the reporting of PA by GPAQ, with greater evidence of accuracy in reporting for those with stable work patterns. Further evidence from this thesis was the seasonal variation in reporting of PA that has been reported previously (10, 59), and that suggests respondents in developing countries are unduly influenced in reporting by their most recent activity. In this population, reporting of fruit and vegetable intake also varied by season. The GPAQ was designed as an improvement on IPAQ, but the results of a recent assessment (60) have shown that, similar to IPAQ, the
GPAQ has only poor to fair criterion validity and moderate reliability. It nonetheless has been claimed to be a suitable and acceptable instrument for monitoring the PA of populations (60). The findings on the issues that arise in the administration of this questionnaire in low- and middle-income countries, such as Vietnam, provide possible explanations for its unfavourable psychometric properties. Partly due to exaggerated PA values in this study, there are reporting issues that arise irrespective of country of application because the PA data are zero-inflated and right-skewed. The GPAQ Analysis Guide (61) provides limited guidance in these respects, however. Of the several methods for handling the zero-inflated and right-skewed data, the Box-Cox transformations produced the most plausible summary values. For data with zero values, this method requires a constant to be added to each observation, and it was possible to choose a value of the constant that produced a design-based mean most like the corresponding median in each stratum and sub-domain. Searching for this value was straightforward and feasible to do.
Information on quantities of fruit and vegetables consumed in terms of standard serving sizes had some evidence of construct validity because the measurements were plausibly correlated with socioeconomic factors in individual-level analyses, and with provincial characteristics in aggregate analyses. However, fruit and vegetable intake was positively associated with body size/fatness, and these associations were independent of age, income, education, tobacco smoking, alcohol consumption, and physical activity. If those who ate more fruit and vegetables were also eating more of other high-energy foods, and total energy intake is a contributor to larger body size/fatness, adjusting for other energy sources would attenuate the positive associations, and could unmask a protective association of fruit/vegetable
consumption with body size/fatness if the cross-correlations with energy intake are of sufficient magnitude (62). Our findings thus suggest that the data collected using the STEPS instrument are not useful for the investigation of the association between fruit and vegetable intake and pathophysiological factors (e.g. BMI, blood glucose and cholesterol) because information on total energy intake is not available.
The four simple questions included in the STEPS questionnaire seeks reporting of fruit and vegetable consumption in a typical week, but the reported fruit and vegetable intake varied by season. Because actual fruit and vegetable consumption patterns display seasonal variability (63), this would be unsurprising were it not for the fact that the STEPS questionnaire asks about comsumption in a typical week. It suggests that Vietnamese people are influenced in
Chapter 7. Summary, implications and future research 156 156 reporting by their intake in the most recent season and, together with seasonal reporting of PA, suggests that a substantial number of respondents found the concept of a typical week difficult to comprehend and deal with. To illustrate the impact of this issue, our data suggest that a survey conducted in the wet season could produce mean estimates of fruit and vegetable intake that are 10% higher than those of a survey conducted in the dry season. Although the seasonal variation in reporting intake in this survey had only minor impact on the national estimates because a large majority of the interviews were conducted in the wet season, this is an important issue to be considered when planning surveys.