 |
Description and quality of baseline data:
|
1 Department of Health Promotion and Chronic Disease Prevention, National Public Health Institute, Helsinki,
Finland
3 Since January 2007 at Indic Society for Education and Development
(INSEED), Nashik, India
3 See Annex for the sites and key
personnel of contributing MORGAM Centres
|
© National Institute for Health and Welfare
and the MORGAM Project investigators Last updated: 4 July 2007 For more information, please contact Kari Kuulasmaa (firstname.lastname@thl.fi) |
MORGAM collected data on height, weight, waist circumference and hip circumference, which were measured in the baseline examination. The data were transferred to the MDC using the Data transfer format - MONICA survey data (Form 20), which has the following relevant data items:
MORGAM also collected an indicator of the waist and hip measurer (item WHCODER). This data item was for the local quality control of the MPCs, and was not used in this assessment.
For data analyses, two derived variables have been defined by the time of preparation of this document:
This description and quality assessment consists of:
The assessment will be done separately for two groups of the measurements:
The MPCs were asked to report the type of scale that was used for the measurement of weight. A traditional balanced beam scale has been considered a reliable instrument for population measurement. Electronic digital scales are easier to operate, but they are usually impossible to calibrate. The accuracy of bathroom scales is insufficient for scientific use.
A Scale Score (SCA) was defined as:
| SCA =
|
2 | if the balance scale was used; |
| 1 | if a digital scale was used; | |
| 0 | if a bathroom scale was used. |
The MPCs were asked to report various details of the procedures, such as clothing of the subjects during the measurements, the resolution to which the measurement values were recorded and whether or not self-reported data were accepted in any situations. For height, for which data to the nearest centimetre was reported for MORGAM, the rounding of the recorded reading, to the full centimetre was asked.
For waist and hip circumference, which are more difficult to measure than weight and height, the general principles of the measurement procedure and training of the measurers were asked.
This involves a review of the measurement distributions and the distributions of the last digits of the measurements. Preference of certain last digits indicates insufficient clarity of the procedures or insufficient training of the measurers. As such problems may be reflected even in the distribution of the second last digits, also this was calculated for weight, and the circumferences of waist and hip.
Following the approach outlined in Quality Assessment of Weight and Height Measurements in the WHO MONICA Project [1] and Quality Assessment of Data on Waist and Hip Circumferences in the WHO MONICA Project [2], quality scores based on the terminal digits were defined.
A subject's measurement was considered complete only if both height and weight were recorded. A Proportion of Incomplete Measurements score (PIM) was defined as:
| PIM = | 2 | if less than 2.5% of the measurements were incomplete; |
| 1 | if 2.5 ≤ proportion of incomplete measurements ≤ 5%; | |
| 0 | if proportion of incomplete measurements > 5%; | |
| dnp | if the data were not provided by the MPC. |
It is reasonable to expect that the terminal digits of height are distributed uniformly, and hence that about 10% of the readings have zero as the terminal digit. If the measurement was not done carefully, the proportion of zeros as the terminal digit can be expected to be larger than 10%.
MONICA suggested cut points 12% and 13% for the proportion of zeros. The cut-off points 13% and 14% used in the MORGAM scores allow slight deviations from the uniform distribution and are more reasonable for RUAs where the sample size was less than 1000 subjects.
The Height Score (HEI ) was defined as:
| HEI = | 2 | if the proportion of terminal zeros is ≤ 13%; |
| 1 | if not 2 or 0; | |
| 0 | if the proportion of terminal zeros is > 14%. |
MORGAM instructions recommended recording the body weight to the nearest 200 g, which is the usual resolution of the beam balance scale. There is no harm if weight is recorded to the nearest 100 g, which is more appropriate than 200 g if the measurement device allows this, as is the case with a digital scale. Therefore, the cut point of good quality was taken at 30%, which indicates a concern over both of these resolutions if there is zero preference or if a resolution of 500g or higher is used. If the measurements are not done properly, there can also be preference of full kilograms. If the terminal digits are uniformly distributed, about 10% of the readings have zero as the second last digit.
The Weight Score (WEI) was defined as:
| WEI = | 2 | if the proportion of terminal zeros is ≤ 30% AND there are ≤ 13% zeros in the second last digit; |
| 1 | if not 2 or 0; | |
| 0 | if the proportion of terminal zeros is > 60% OR there are > 14% zeros in the second last digit. |
The Summary Score of Measurements (SSM) for height and weight was derived using the Scale Score, Height Score and Weight Score:
| SSM = | 2 | if the sum is 5 or 6; |
| 1 | if the sum is 3 or 4; | |
| 0 | if the sum is 0, 1 or 2. |
The Summary Score of Quality (SSQ) was defined to summarize the quality taking into account the proportion of incomplete data. It was defined in a such way that SSQ will become smaller than SSM if PIM is less than two, i.e.
|
SSQ = max(0, SSM - (2-PIM)). |
A Proportion of Incomplete Measurements score (PIM) for waist and hip measurements was defined as for height and weight measurements.
MORGAM instructions recommended recording the waist and hip circumference to centimetres and a decimal, rounded to the nearest 0 or 5. If the terminal digits are uniformly distributed, about 50% of the readings have zero as the last digit. When looking at the distributions of terminal digits in Table 4a and Table 5a, it is evident that the zero preferences in waist and hip measurements are very similar. This is understandable since both measurements are done at the same time and in the same way by the same measurer. Digit preference in the measurements indicates insufficient training of the measurers. If the data are of good quality, the proportion of terminal zeros is expected to be between 30% and 70%. These limits are also exceeded if the rounding was done to the nearest full centimetre. If the measurements are not done properly, there can also be preference of full centimetres. Similar cut-off points as for height measurement are used for full centimetres of waist and hip measurements in MORGAM.
The Waist Score (WAI) and the Hip Score (HIP) were defined as:
| WAI/HIP = | 2 | if the proportion of terminal zeros is between 30% and
70% AND there are ≤ 13% zeros in the second last digit; |
| 1 | if not 2 or 0; | |
| 0 | if there are >14% zeros in the second last digit. |
The score gives little weight on zero-preference of the decimal (i.e. the last digit) because the measurement of waist and hip circumference to that precision is difficult.
The Summary Score of Measurements (SSM) for waist and hip was derived using the sum of waist and hip scores:
| SSM = | 2 | if the sum is 3 or 4; |
| 1 | if Waist Score=1 and Hip Score=1; | |
| 0 | if Waist Score=0 or Hip Score=0. |
The Summary Score of Quality (SSQ) was defined to summarize the quality taking into account the proportion of incomplete data. It was defined in a such way that SSQ will become smaller than SSM if PIM is less than two, i.e.
|
SSQ = max(0, SSM - (2-PIM)). |
Table1a reports the type of scale used. In most RUAs, balance scales were used in all cohorts. Digital scales were used in FIN-ATB, in the PRIME Cohorts (FRA-LIL/STR/TOU and UNK-BEL) and in all Scottish Cohorts (UNK-EDI/GLA/SHH). In Cohort 01 and 02 of GER-AUG, bathroom scales were used for some members of the cohorts examined on home visits, otherwise balance scales were used.
Shoes and heavy outer garments were removed before the measurements of weight and height in all Cohorts.
Following the WHO MONICA Procedures [3], the recordings of height were transferred to the MDC in full centimetres. The resolution of recording weight locally in the RUA and the rounding of the values for transfer to the MDC are shown in Table1a. Most MPCs have gathered their data using the resolution of 1 or 0.1 cm. For the Cohorts, where the resolution was locally 0.5 cm, the rounding for MORGAM becomes critical. The resolution was 0.5 cm in DEN-GLO, GER-AUG, ITA-FRI/FSE, ITA-ROM (Cohorts 23 and 24), LTU-KAU (Cohorts 01 and 02), POL-TAR and POL-WAR. For estimating the mean height of the Cohort, the direction of the rounding is crucial. LTU-KAU made the rounding down, all others made it up. We do not know the direction of the rounding in ITA-ROM. For follow-up analysis, which is based on the variation between individuals n the cohorts, as in most of the MORGAM analysis, it is most important that rounding is always done in the same direction. This is the case with all cohorts, although there is a change from Cohort 02 to Cohort 03 in LTU-KAU, which needs to be taken into account in the data analysis.
Table1a shows the reported resolution of weight measurement in each cohort. Nearly all centres reported having recorded weight to the nearest 100 g or 200g. ITA-ROM reported recording to full kilograms and UNK-EDI/GLA/SHH in Cohort 01 to the nearest 500 g. Table 3a gives the distribution of the decimals of kilograms of the actual data. Most MPCs have provided the data on weight with the resolution which they have also reported having used locally. However, there are many cohorts which have transferred the data on 200 g although they have reported using locally 100 g resolution. This is easier to understand than the situation in the many cohorts which have transferred data to 100 g but said that it was reported locally to the nearest 200 g. Cohorts with significant findings are:
Self-reported data instead of measuring height and weight were accepted for non-ambulatory subjects (or for ambulatory subjects who are in such a way disabled that could not be measured) in Cohorts 03 and 24 of FIN-EAS/WES, in ITA-FRI/FSE, in SWE-NSW and in UNK-EDI/GLA/SHH. Such cases are rare. In Cohorts 24 of FIN-EAS/WES, self-reported data were accepted also for other subjects examined during home visits, which is 2% of all subjects. Nearly all of them are from age group 65-74 years, where they comprise 17% in FIN-EAS and 11% in FIN-WES.
MORGAM did not collect the data on the quality assurance of the height and weight measurements in the MPCs.
Hyperlinks to the distributions of the data items are under the respective data item names:
The most striking finding from the distributions of the height and weight data is that the proportion of missing values is negligible. The highest proportion of missing values is 3.2% in Cohort 02 of RUS-NOV.
Table 2 gives the distributions of the terminal digits for height in the different Cohorts. The main finding in Table 2 is that overall, the distributions are very uniform. The percentage of terminal zeros is only rarely above the critical values.
Table 3a gives the distributions of the terminal digits (i.e. the decimals of kg) for weight measurement recorded in the MORGAM Cohorts. The proportion of terminal zeros is > 60% only in the cohorts which reported using full kilograms, or where only full kilograms were transferred to the MDC. In addition, the proportion of terminal zeros is above 30% in a number of cohorts which reported having measured weight to the nearest 100 g or 200 g.
Table 3b gives the distributions of the second last digit (i.e. the terminal digits for the full kilograms). The proportion of terminal zeros is over 13% in ITA-FSE and in Cohort 03 of LTU-KAU only.
Table 6 summarizes the quality of height and weight measurements. The Summary Score of Quality is less than two for:
For these Cohorts there is indication of insufficient training of the measurers and/or insufficient quality control of the measurements.
All MPCs, that provided the data on waist and hip, reported having measured waist and hip circumferences according to the MONICA instructions [3].
Table 1b shows the reported procedures used for waist and hip measurements. The measurers were trained in all MPCs and tested or certified in all except DEN-GLO, FRA-LIL/STR, POL-WAR and SWE-NSW.
The MONICA instructions recommended to measure waist and hip circumference on subjects while they are semi-clothed, i.e. waist uncovered with the subjects wearing underclothes only. If it is not possible, the instructions advise to measure the circumferences on subjects without heavy outer garments with all tight clothing, including the belt, loosened and with the pockets emptied. Accordingly, removal of heavy garments before measurements were done in all Centres, and removal of other garments were done in all Centres except AUS-NEW, DEN-GLO, FIN-EAS/WES, FRA-TOU, UNK-BEL and UNK-GLA. Each subject was measured by one person in all Cohorts except Cohort 03 of AUS-NEW, where two persons were involved.
Table 1b shows the reported resolution of waist and hip measurements in each cohort. AUS-NEW and DEN-GLO reported recording to full centimetres. Table 4a and Table 5a give the distribution of the decimals of centimetres of the actual data for waist and hip respectively. In Cohort 03 of AUS-NEW, the data are provided in centimetres and a decimal, which contradicts the procedures reported by the MPC in Table 1b. Other decimals than 0 or 5 were recorded for occasional subjects in FIN-WES, LTU-KAU, POL-TAR and POL-WAR, and more frequently in:
Hyperlinks to the distributions of the data items are under the respective data item names:
The data on waist and hip circumference are available for about a half of the cohorts. The data are not available from FIN-ATB, RUS-NOV, UNK-CAE, UNK-EDI and UNK-SHH. From other Centres the data are available at least for the latest Cohort.
The proportion of missing values is very small in Cohorts where the data were collected. The highest proportions of the missing data for waist and hip circumference are 4-5% in Cohorts 02 and 21 of ITA-BRI/PAM.
Table 4a and Table 5a give the distributions of the terminal digits for waist and hip circumference respectively in decimals in the MORGAM Cohorts. In addition to the cohorts recording only full centimetres, there is a tendency toward full centimetres also in all other cohorts. The proportion of terminal zeros is over 70% in a half of the cohorts and over 50% in all cohorts both for waist and hip measurements.
Table 4b and Table 5b give the distributions of the terminal digits for waist and hip circumference in full centimetres, respectively. The proportion of terminal zeros is over 13% in Cohort 02 of ITA-BRI only.
Table 7 summarizes the quality of waist and hip data. The Summary Score of Quality is less than two for:
In particular for Cohort 02 of ITA-BRI and ITA-PAM, where the summary score is "0", there is indication of insufficient training of the measurers and/or insufficient quality control of the measurements.
| Date | Update |
|---|---|
| 2007-07-04 | First published version. |