EUROPEAN HEALTH INTERVIEW URVEY - circabc.europa.eu wave 2... · • SC = Skip checks, which follows the filters defined in the annex of the EHIS Regulation; • CC = Consistency

EUROPEAN COMMISSION EUROSTAT Directorate F: Social statistics Unit F-5: Education, health and social protection

EEUURROOPPEEAANN HHEEAALLTTHH IINNTTEERRVVIIEEWW SSUURRVVEEYY

WWAAVVEE 22

VVAALLIIDDAATTIIOONN RRUULLEESS

Version 23 November 2015

- 2 -

1. INTRODUCTION

The Commission regulation for the European Health Interview Survey (EHIS)1 specifies in article 6 that: “Member States shall provide finalised, validated and weighted microdata and quality-related reference metadata required by this Regulation in accordance with an exchange standard specified by the Commission (Eurostat)”. The annex specifies for each variable the title or short definition, the codename, the answer codes and values and the filters.

In addition, the Commission implementing decision on derogations2 specifies for three listed countries a set of variables which have not to be delivered.

This document describes the validation rules to be applied to the standard microdata as defined by the Regulation. Member States should apply this validation on the dataset – specified in the data delivery guidelines - before sending it to Eurostat.

Obviously Member States may need to use different and/or complementary validation rules on their national data if their questionnaire differs from the model questionnaire proposed in the methodological manual.

More information on the EHIS wave 2 methodology and the description of variables and model questions can be found in the European Health Interview Survey (EHIS wave 2) methodological manual3.

2. VALIDATION RULES

Three groups of validation rules are distinguished:

• VC = Checks for allowed codes and values, including allowed ranges; these rules reflect the definitions of the variables in the annex of the EHIS Regulation;

• SC = Skip checks, which follows the filters defined in the annex of the EHIS Regulation;

• CC = Consistency checks (relations between variables from the same sub-module or between variables from different (sub-)modules).

Validation rules result in indicating:

• E = Error: Problems in data that should be solved and should not appear in the final microdata file; or

• W = Warning: Indicating some unusual values or combination of values of variables which may exist but should not be highly prevalent. The validation report will in its summary give for each rule the prevalence for failing. The validation process should not be stopped for warnings but a high prevalence for a warning validation rule might pinpoint to the responsible statistician a possible structural problem.

1 Commission Regulation (EU) No 141/2013 of 19 February 2013 implementing Regulation (EC) No 1338/2008 of

the European Parliament and of the Council on Community statistics on public health and health and safety at work, as regards statistics based on the European Health Interview Survey (EHIS).

2 Commission Implementing Decision of 19 February 2013 granting derogations to certain Member States with respect to the transmission of statistics pursuant to Regulation (EC) No 1338/2008 of the European Parliament and of the Council on Community statistics on public health and health and safety at work, as regards statistics based on the European Health Interview Survey (EHIS).

3 http://ec.europa.eu/eurostat/en/web/products-manuals-and-guidelines/-/KS-RA-13-018

http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2013:047:0020:0048:EN:PDF






http://ec.europa.eu/eurostat/en/web/products-manuals-and-guidelines/-/KS-RA-13-018

- 3 -

All rules resulting in an error and the rules resulting in a warning for which we have marked in the table below a note ‘The records marked with the warning are recommended to be checked all’ should ideally be incorporated as immediate checks during the interview.

Eurostat validation tool will include some of the following logical verifications previously considered to be applied only at national level. Therefore, they will result only in warning, listing records which do not comply with the rule. These rules are:

Reported item Rule

No duplication of PID (unique identifier) Each PID code is unique.

No duplication of HHID for different households (unique household identifier). For one value of HHID only the same combination of PSU, PRIMSTRAT, COUNTRY, REGION, DEG_URB, HHNBPERS, HHTYPE, HHINCOME must be recorded.

For all records with the same value (but different from -1) of HHID only unique values of the following variables are possible: PSU, PRIMSTRAT, COUNTRY, REGION, DEG_URB, HHNBPERS, HHNBPERS_0_4, HHNBPERS_5_13, HHNBPERS_14_15, HHNBPERS_16_24, HHNBPERS_25_64, HHNBPERS_65plus, HHTYPE, HH_ACT, HH_INACT, HHINCOME

Consistency of household data for people from one household (if more than one person from a household is interviewed). This concerns variables which describe or reflect relations of persons in household, for example marital status or informal care.

The rule is not implemented in the Eurostat validation tool.

Unique link between PSU and PRIMSTRAT.

Only for records for which (PSU ≠ -2 and PRIMSTRAT ≠ -2): for records with the same value of PSU the code of PRIMSTRAT is the same.

Running the validation will issue reports. Eurostat validation tool will include some of the additional features in such reporting but we prefer to report and follow those up in the quality reports:

Reported item Reported indicator

The share of interviews which cannot be treated as completed as at least one of the key variables includes the missing value

Percentage of records with at least one value equal to ‘-1’ in at least one of the following variables: PID, HHID, PRIMSTRAT, PSU, WGT, PROXY, REFYEAR, REFMONTH, INTMETHOD, SEX and AGE.

The share of incomplete interviews (plain) Percentage of records for which: a) at least one of the following variables include a missing value (NULL or ‘-1’): PID, HHID, PRIMSTRAT, PSU, WGT, PROXY, REFYEAR, REFMONTH, INTMETHOD, SEX and AGE, or: b) more than 50 % of all other variables in the record include missing value (NULL or ‘-1’). The denominator is the same for each record and is equal to the total number of variables – 11 = 151 – 11 = 140.

- 4 -

The share of incomplete interviews (derogation and proxy-adjusted)

Percentage of records for which: a) at least one of the following variables include a missing value (NULL or ‘-1’): PID, HHID, PRIMSTRAT, PSU, WGT, PROXY, REFYEAR, REFMONTH, INTMETHOD, SEX and AGE, or: b) more than 50 % of all other variables in the record (relevant to the specific interview) include missing value (NULL or ‘-1’). The numerator and denominator need to be adjusted to remove variables for which proxy interviews are not allowed provided that the record include (PROXY = 2 OR 3) and variables for which derogations were granted (see the list of countries and respective variables at the end of this document).

The share of proxy interviews Percentage of records with the values of PROXY variable equal to any of 2 or 3. Records with values ‘-1’ are excluded from denominator.

REFMONTH is September-December at least for some records (EHIS Regulation, art. 4, par. 3)

Percentage of records with the values of REFMONTH variable equal to any of 9, 10, 11 or 12. Records with values ‘-1’ are excluded from denominator.

Extreme weights (any weight is more than 3-times higher/lower than average weight) (for WGT)

Percentage of records with the value of WGT variable out of the following range: < average value of WGT / 3; 3 * average value of WGT >. Records with values ‘-1’ are excluded from denominator.

The share of missing values (item non-response)

Percentage of records with the value ‘-1’ calculated for each variable.

Frequencies of income quintile groups (HHINCOME) are similar, i.e. each of the categories has frequency of about 20%.

The frequency of each of the five income quintile group categories is between 15 and 25%.

Only records with values 1, 2, 3, 4, 5 are included in the calculation.

1. Checks for allowed values (VC):

Other codes than stated in the following table are to be treated as errors (E).

Official classifications and standard code lists are available on http://ec.europa.eu/eurostat/ramon/.

Code Allowed codes / range Note VC_PID -1; 1 - 9999999999 VC_HHID -1; 1 - 9999999999 VC_PRIMSTRAT -2; 0001 - 9999 Code -1 is allowed but its use would issue an error message VC_PSU -2; 0001 - 9999 Code -1 is allowed but its use would issue an error message VC_WGT -1; 0.001 – 99999.999 8 digits in total – maximum 5 digits for the integer part and 3 digits

for the fractional part VC_PROXY -1; 1; 2; 3 VC_REFYEAR 2013; 2014; 2015 Code -1 is allowed but its use would issue an error message VC_REFMONTH -1; 1 - 12 VC_INTMETHOD 10; 11; 20; 21; 30; 31; 40; 50 Code -1 is allowed but its use would issue an error message VC_INTLANG -1;

BUL; SPA; CZE; DAN; GER; EST; GRE; ENG; FRE; GLE; ITA; LAV; LIT; HUN; MLT; DUT; POL; POR; RUM; SLO; SLV; FIN; SWE; BAQ; CTL; GLG; LTZ; VAL; WEL; NOR; HRV; ICE; MAC; SRP; TUR; ALB; BEL; BOS; RUS; UKR; ARA; CHI; HIN; IND; JPN; KOR; MON; NEP; VIE; OTH

SCL - Languages

VC_SEX -1; 1; 2 The value ‘-1’ is theoretically possible but such record will be treated as non-response.

VC_AGE -1; 15 - 120 The value ‘-1’ is theoretically possible but such record will be treated as non-response.

VC_COUNTRY -1; BE; BG; CZ; DK; DE; EE; IE; EL; ES; FR; HR; IT; CY; LV; LT; LU; HU; MT; NL; AT; PL; PT; RO; SI; SK; FI;

SCL - Geographical code list (NUTS country level) Only EU 28, EFTA and EU candidate countries in the list; code ZZ

http://ec.europa.eu/eurostat/ramon/

- 6 -

Code Allowed codes / range Note SE; UK; IS; LI; NO; CH; ME, MK, AL, RS, TR; ZZ

for other countries was added (but specific codes for other countries may also be added). For each country only the relevant country code should be used (plus code '-1' for missing records), i.e. for Belgium only 'BE' and '-1' are the allowed codes.

VC_BIRTHPLACE -1; 10; 21; 22 VC_CITIZEN -1; 10; 21; 22 VC_REGION -1; 10 – ZZ

NUTS level 2 of the NUTS classification (Standard Code List): http://ec.europa.eu/eurostat/ramon/nomenclatures/index.cfm?TargetUrl=LST_NOM_DTL&StrNom=NUTS_33&StrLanguageCode=EN&IntPcKey=&StrLayoutCode=HIERARCHIC&IntCurrentPage=1 The requested code refers to the 2 digits on the 3rd and 4th position of the NUTS2 code (NUTS 2010 classification). Codes to be used are country-specific (relation to/compatibility with COUNTRY variable should be ensured).

VC_DEG_URB -1; 1; 2; 3 VC_MARSTALEGAL -1; 1; 2; 3; 4 VC_MARSTADEFACTO -1; 1; 2 VC_HATLEVEL -1; 0; 1; 2; 3; 4; 5; 6; 7; 8 VC_MAINSTAT -1; 10; 20; 31; 32; 33; 34; 35; 36 VC_FT_PT -2; -1; 1; 2 VC_JOBSTAT -2; -1; 10; 21; 22 VC_JOBISCO -2; -1;

11; 12; 13; 14; 21; 22; 23; 24; 25; 26; 31; 32; 33; 34; 35; 41; 42; 43; 44; 51; 52; 53; 54; 61; 62; 63; 71; 72; 73; 74; 75; 81; 82; 83; 91; 92; 93; 94; 95; 96; 0

SCL - International Standard Classification of Occupations 2008 (ISCO-08). Code 0 used for 'Armed forces occupations'.

VC_LOCNACE -2; -1; A - U

SCL - Statistical Classification of Economic Activities in the European Community (NACE Rev. 2) (sections)

http://ec.europa.eu/eurostat/ramon/nomenclatures/index.cfm?TargetUrl=LST_NOM_DTL&StrNom=NUTS_33&StrLanguageCode=EN&IntPcKey=&StrLayoutCode=HIERARCHIC&IntCurrentPage=1




- 7 -

Code Allowed codes / range Note VC_HHNBPERS -1; 1 - 98 VC_HHNBPERS_0_4 -1; 0 - 98 VC_HHNBPERS_5_13 -1; 0 - 98 VC_HHNBPERS_14_15 -1; 0 - 98 VC_HHNBPERS_16_24 -1; 0 - 98 VC_HHNBPERS_25_64 -1; 0 - 98 VC_HHNBPERS_65plus -1; 0 - 98 VC_HHTYPE -1; 10; 21; 22; 23; 24; 25 VC_HH_ACT -1; 0 - 98 VC_HH_INACT -1; 0 - 98 VC_HHINCOME -1; 1; 2; 3; 4; 5 VC_HS1 -1; 1; 2; 3; 4; 5 VC_HS2 -1; 1; 2 VC_HS3 -1; 1; 2; 3 VC_CD1A -1; 1; 2 VC_CD1B -1; 1; 2 VC_CD1C -1; 1; 2 VC_CD1D -1; 1; 2 VC_CD1E -1; 1; 2 VC_CD1F -1; 1; 2 VC_CD1G -1; 1; 2 VC_CD1H -1; 1; 2 VC_CD1I -1; 1; 2 VC_CD1J -1; 1; 2 VC_CD1K -1; 1; 2 VC_CD1L -1; 1; 2 VC_CD1M -1; 1; 2 VC_CD1N -1; 1; 2 VC_CD1O -1; 1; 2

- 8 -

Code Allowed codes / range Note VC_AC1A -1; 1; 2 VC_AC1B -1; 1; 2 VC_AC1C -1; 1; 2 VC_AC2 -2; -1; 1; 2; 3 Adaptation of the rule is needed for preliminary values checks in

case the answer categories proposed in the model questionnaire are used: -2; -1; 1; 2; 3; 4

VC_AW1 -2; -1; 1; 2 VC_AW2 -2; -1; 1 - 365 VC_PL1 -1; 1; 2; 3 VC_PL2 -2; -1; 1; 2; 3; 4 VC_PL3 -1; 1; 2; 3 VC_PL4 -2; -1; 1; 2; 3; 4 VC_PL5 -2; -1; 1; 2; 3; 4 VC_PL6 -1; 1; 2; 3; 4 VC_PL7 -1; 1; 2; 3; 4 VC_PC1A -2; -1; 1; 2; 3; 4 VC_PC1B -2; -1; 1; 2; 3; 4 VC_PC1C -2; -1; 1; 2; 3; 4 VC_PC1D -2; -1; 1; 2; 3; 4 VC_PC1E -2; -1; 1; 2; 3; 4 VC_PC2 -2; -1; 1; 2 VC_PC3 -2; -1; 1; 2 VC_HA1A -2; -1; 1; 2; 3; 4; 5 VC_HA1B -2; -1; 1; 2; 3; 4; 5 VC_HA1C -2; -1; 1; 2; 3; 4; 5 VC_HA1D -2; -1; 1; 2; 3; 4; 5 VC_HA1E -2; -1; 1; 2; 3; 4; 5 VC_HA1F -2; -1; 1; 2; 3; 4; 5 VC_HA1G -2; -1; 1; 2; 3; 4; 5

- 9 -

Code Allowed codes / range Note VC_HA2 -2; -1; 1; 2 VC_HA3 -2; -1; 1; 2 VC_PN1 -1; 1; 2; 3; 4; 5; 6 VC_PN2 -1; 1; 2; 3; 4; 5 VC_MH1A -1; 1; 2; 3; 4 VC_MH1B -1; 1; 2; 3; 4 VC_MH1C -1; 1; 2; 3; 4 VC_MH1D -1; 1; 2; 3; 4 VC_MH1E -1; 1; 2; 3; 4 VC_MH1F -1; 1; 2; 3; 4 VC_MH1G -1; 1; 2; 3; 4 VC_MH1H -1; 1; 2; 3; 4 VC_HO1 -1; 1; 2 VC_HO2 -2; -1; 1 - 365 The EHIS Regulation refers to codes -2, -1; 1-99 but codes 100 –

365 are acceptable. VC_HO3 -1; 1; 2 VC_HO4 -2; -1; 1 - 365 The EHIS Regulation refers to codes -2, -1; 1-300 but codes 301 –

365are acceptable. VC_AM1 -1; 1; 2; 3; 4 VC_AM2 -1; 1; 2; 3 VC_AM3 -2; -1; 0 - 99 VC_AM4 -1; 1; 2; 3 VC_AM5 -2; -1; 0 - 99 VC_AM6A -1; 1; 2 VC_AM6B -1; 1; 2 VC_AM7 -1; 1; 2 VC_MD1 -1; 1; 2 VC_MD2 -1; 1; 2 VC_PA1 -1; 0; MMYYYY MM (month): 01 – 12; YYYY (year): 2010 – 2015. 4-digit code

- 10 -

Code Allowed codes / range Note YYYY (year) is allowed in case MM (month) is missing. The rule applies to data in the format for data delivery to Eurostat as requested by the EHIS Regulation. Adaptation of the rule is needed for preliminary values checks in case the answer categories proposed in the model questionnaire are used: -1; 1; 2; MMYYYY

VC_PA2 -1; 1; 2; 3; 4; 5 VC_PA3 -1; 1; 2; 3; 4; 5 VC_PA4 -1; 1; 2; 3; 4; 5 VC_PA5 -1; 1; 2; 3; 4; 5 VC_PA6 -1; 1; 2; 3; 4; 5 VC_PA7 -2; -1; 1; 2; 3; 4; 5 VC_PA8 -2; -1; 1; 2; 3; 4; 5 VC_UN1A -1; 1; 2; 3 VC_UN1B -1; 1; 2; 3 VC_UN2A -1; 1; 2; 3 VC_UN2B -1; 1; 2; 3 VC_UN2C -1; 1; 2; 3 VC_UN2D -1; 1; 2; 3 VC_BM1 -1; 50 - 250 VC_BM2 -1; 20 – 350 VC_PE1 -1; 1; 2; 3; 4 VC_PE2 -1; 0 - 7 VC_PE3 -2; -1; 1; 2; 3; 4; 5 VC_PE4 -1; 0 - 7 VC_PE5 -2; -1; 1; 2; 3; 4; 5 VC_PE6 -1; 0 - 7

- 11 -

Code Allowed codes / range Note VC_PE7 -2; -1; HHMM HH (hours): 00 – 99; MM (minutes): 00 – 59. All four digits shall

be reported, i.e. 0020 for 20 min or 0900 for 9 hours. VC_PE8 -1; 0 - 7 VC_FV1 -1; 1; 2; 3; 4; 5 VC_FV2 -2; -1; 1 - 99 VC_FV3 -1; 1; 2; 3; 4; 5 VC_FV4 -2; -1; 1 - 99 VC_SK1 -1; 1; 2; 3 VC_SK2 -2; -1; 1; 2; 3; 4 VC_SK3 -2; -1; 1 - 99 VC_SK4 -1; 1; 2; 3 VC_AL1 -1; 1; 2; 3; 4; 5; 6; 7; 8; 9 VC_AL2 -2; -1; 1; 2; 3; 4; 5 VC_AL3 -2; -1; 1; 2; 3; 4; 5; 6; 7; 8 VC_AL4 -2; -1; 1; 2; 3; 4 VC_AL5 -2; -1; 1; 2; 3; 4; 5; 6; 7; 8 VC_AL6 -2; -1; 1; 2; 3; 4; 5; 6; 7; 8; 9 VC_SS1 -1; 1; 2; 3; 4 VC_SS2 -1; 1; 2; 3; 4; 5 VC_SS3 -1; 1; 2; 3; 4; 5 VC_IC1 -1; 1; 2 VC_IC2 -2; -1; 1; 2 VC_IC3 -2; -1; 1; 2; 3

Warnings on values

Code Rule (indication of a problem) Type Notes VC_HHNBPERS_W HHNBPERS > 12 W The records marked with the warning are

recommended to be checked all.

- 12 -

VC_AM3_W AM3 > 20 W VC_AM5_W AM5 > 20 W VC_BM1_W (BM1 > 0 and BM1 < 70) or (BM1 > 220) W VC_BM2_W BM2 > 150 W VC_FV2_W FV2 > 20 W VC_FV4_W FV4 > 20 W

- 13 -

2. Skip checks (filters in the questionnaire) (SC)

Code Rule (indication of a problem) Type Description of/comment on the rule

Notes

SC_FT_PT (MAINSTAT ≠ 10) and (FT_PT ≠ -2) E FT_PT should be '-2' or MAINSTAT equal to '10'.

If both values are -1, FT_PT should be coded as -2

SC_JOBSTAT (MAINSTAT ≠ 10) and (JOBSTAT ≠ -2) E JOBSTAT should be '-2' or MAINSTAT equal to '10'.

If both values are -1, JOBSTAT should be coded as -2

SC_JOBISCO (MAINSTAT ≠ 10) and (JOBISCO ≠ -2) E JOBISCO should be '-2' or MAINSTAT equal to '10'.

If both values are -1, JOBISCO should be coded as -2

SC_LOCNACE (MAINSTAT ≠ 10) and (LOCNACE ≠ -2) E LOCNACE should be '-2' or MAINSTAT equal to '10'.

If both values are -1, LOCNACE should be coded as -2

SC_AC2 (AC1A ≠ 1 and AC1B ≠ 1 and AC1C ≠ 1) and (AC2 ≠ -2)

E AC2 should be '-2' or (AC1A = 1 or AC1B = 1 or AC1C = 1).

SC_AW1 (MAINSTAT ≠ 10) and (AW1 ≠ -2) E AW1 should be '-2' or MAINSTAT equal to '10'.

If both values are -1, AW1 should be coded as -2

SC_AW2 (AW1 ≠ 1) and (AW2 ≠ -2 or -1) E AW1 should be 1 or AW2 '-2', resp. '-1' (proxy).

SC_PL2 (PL1 = 3) and (PL2 ≠ -2 or -1) E PL1 should not be 3 or PL2 equal to '-2'.

The rule was made less strict, allowing answering PL2 in case PL1 is missing.





SC_PL45 (PL4 = 4) and (PL5 ≠ -2) E If PL4 = 4 then PL5 should be equal to '-2'.

- 14 -


Notes

SC_PC2 {(PC1A ≠ 2 or 3 or 4) and (PC1B ≠ 2 or 3 or 4) and (PC1C ≠ 2 or 3 or 4) and (PC1D ≠ 2 or 3 or 4) and (PC1E ≠ 2 or 3 or 4)} and (PC2 ≠ -2)

E PC2 should be filled in only if at least one of PC1A to PC1E is different from 1, -1 and -2).

SC_PC3 {(PC1A ≠ 2 or 3 or 4) and (PC1B ≠ 2 or 3 or 4) and (PC1C ≠ 2 or 3 or 4) and (PC1D ≠ 2 or 3 or 4) and (PC1E ≠ 2 or 3 or 4)} and (PC3 ≠ -2 or -1)

E PC3 should be filled in only if at least one of PC1A to PC1E is different from 1, resp. '-1' (proxy).

SC_HA2 {(HA1A ≠ 2 or 3 or 4) and (HA1B ≠ 2 or 3 or 4) and (HA1C ≠ 2 or 3 or 4) and (HA1D ≠ 2 or 3 or 4) and (HA1E ≠ 2 or 3 or 4) and (HA1F ≠ 2 or 3 or 4) and (HA1G ≠ 2 or 3 or 4)} and (HA2 ≠ -2)

E HA2 should be filled in only if at least one of HA1A to HA1G is different from 1, 5, -1 and -2.

SC_HA3 {(HA1A ≠ 2 or 3 or 4) and (HA1B ≠ 2 or 3 or 4) and (HA1C ≠ 2 or 3 or 4) and (HA1D ≠ 2 or 3 or 4) and (HA1E ≠ 2 or 3 or 4) and (HA1F ≠ 2 or 3 or 4) and (HA1G ≠ 2 or 3 or 4)} and (HA3 ≠ -2 or -1)

E HA3 should be filled in only if at least one of HA1A to HA1G is different from 1, 5, -1 and -2, resp. '-1' (proxy).

SC_HO2 (HO1 ≠ 1) and (HO2 ≠ -2) E HO1 should be 1 or HO '-2'. SC_HO4 (HO3 ≠ 1) and (HO4 ≠ -2) E HO3 should be 1 or HO4 '-2'. SC_AM3 (AM2 ≠ 1) and (AM3 ≠ -2) E AM2 should be 1 or AM3 '-2'. SC_AM5 (AM4 ≠ 1) and (AM5 ≠ -2) E AM4 should be 1 or AM5 '-2'. SC_PA7 (SEX ≠ 2) and (PA7 ≠ -2 or -1) E SEX should be 2 or PA7 '-2',

resp. '-1' (proxy).

SC_PA8 (SEX ≠ 2) and (PA8 ≠ -2 or -1) E SEX should be 2 or PA8 '-2', resp. '-1' (proxy).

SC_PE3 (PE2 ≠ 1 or 2 or 3 or 4 or 5 or 6 or 7) and (PE3 ≠ -2 or -1)

E PE2 should be 1 to 7 or PE3 equal to '-2', resp. '-1' (proxy).



- 15 -


Notes



SC_FV2 (FV1 ≠ 1) and (FV2 ≠ -2 or -1) E FV1 should be 1 or FV2 equal to '-2', resp. '-1' (proxy).

SC_FV4 (FV3 ≠ 1) and (FV4 ≠ -2 or -1) E FV3 should be 1 or FV4 equal to '-2', resp. '-1' (proxy).

SC_SK2 (SK1 ≠ 1 or 2) and (SK2 ≠ -2) E SK1 should be 1 or 2 or SK2 equal to '-2'.

SC_SK3 {(SK1 ≠ 1) or (SK2 ≠ 1)} and (SK3 ≠ -2) E SK1 and SK2 should be 1; or SK3 equal to '-2'.

SC_AL2 (AL1 ≠ 1 or 2 or 3 or 4) and (AL2 ≠ -2 or -1) E AL1 should be 1 to 4 or AL2 equal to '-2', resp. '-1' (proxy).

SC_AL3 (AL2 ≠ 1 or 2 or 3 or 4) and (AL3 ≠ -2 or -1) E AL2 should be 1 to 4; or AL3 equal to '-2', resp. '-1' (proxy).

SC_AL4 (AL1 ≠ 1 or 2 or 3 or 4) and (AL4 ≠ -2 or -1) E AL1 should be 1 to 4 or AL4 equal to '-2', resp. '-1' (proxy).

SC_AL5 (AL4 ≠ 1 or 2 or 3) and (AL5 ≠ -2 or -1) E AL4 should be 1 to 3; or AL5 equal to '-2', resp. '-1' (proxy).

SC_AL6 (AL1 ≠ 1 or 2 or 3 or 4 or 5 or 6 or 7) and (AL6 ≠ -2 or -1)

E AL1 should be 1 to 7 or AL6 equal to '-2', resp. '-1' (proxy).

SC_IC2 (IC1 ≠ 1) and (IC2 ≠ -2 or -1) E IC1 should be 1 or IC2 equal to '-2', resp. '-1' (proxy).

SC_IC3 (IC1 ≠ 1) and (IC3 ≠ -2 or -1) E IC1 should be 1 or IC3 equal to '-2', resp. '-1' (proxy).

- 16 -

3. Consistency checks (CC)

Code Rule Type Description CC_P_HS1 (PROXY = 2 or 3) and (HS1 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_AW2 (PROXY = 2 or 3) and (AW2 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PC3 (PROXY = 2 or 3) and (PC3 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_HA3 (PROXY = 2 or 3) and (HA3 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PN1 (PROXY = 2 or 3) and (PN1 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PN2 (PROXY = 2 or 3) and (PN2 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_MH1A (PROXY = 2 or 3) and (MH1A ≠ -

1) W The variable is not recommended to be filled in by a proxy respondent.

CC_P_MH1B (PROXY = 2 or 3) and (MH1B ≠ -1)

W The variable is not recommended to be filled in by a proxy respondent.

CC_P_MH1C (PROXY = 2 or 3) and (MH1C ≠ -1)


CC_P_MH1D (PROXY = 2 or 3) and (MH1D ≠ -1)


CC_P_MH1E (PROXY = 2 or 3) and (MH1E ≠ -1)


CC_P_MH1F (PROXY = 2 or 3) and (MH1F ≠ -1)


CC_P_MH1G (PROXY = 2 or 3) and (MH1G ≠ -1)


CC_P_MH1H (PROXY = 2 or 3) and (MH1H ≠ -1)


CC_P_MD1 (PROXY = 2 or 3) and (MD1 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_MD2 (PROXY = 2 or 3) and (MD2 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PA1 (PROXY = 2 or 3) and (PA1 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PA2 (PROXY = 2 or 3) and (PA2 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PA3 (PROXY = 2 or 3) and (PA3 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PA4 (PROXY = 2 or 3) and (PA4 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent.

- 17 -

Code Rule Type Description CC_P_PA5 (PROXY = 2 or 3) and (PA5 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PA6 (PROXY = 2 or 3) and (PA6 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PA7 (PROXY = 2 or 3) and (PA7 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PA8 (PROXY = 2 or 3) and (PA8 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_UN1A (PROXY = 2 or 3) and (UN1A ≠ -

1) W The variable is not recommended to be filled in by a proxy respondent.

CC_P_UN1B (PROXY = 2 or 3) and (UN1B ≠ -1)


CC_P_UN2A (PROXY = 2 or 3) and (UN2A ≠ -1)


CC_P_UN2B (PROXY = 2 or 3) and (UN2B ≠ -1)


CC_P_UN2C (PROXY = 2 or 3) and (UN2C ≠ -1)


CC_P_UN2D (PROXY = 2 or 3) and (UN2D ≠ -1)


CC_P_BM1 (PROXY = 2 or 3) and (BM1 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_BM2 (PROXY = 2 or 3) and (BM2 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PE1 (PROXY = 2 or 3) and (PE1 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PE2 (PROXY = 2 or 3) and (PE2 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PE3 (PROXY = 2 or 3) and (PE3 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PE4 (PROXY = 2 or 3) and (PE4 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PE5 (PROXY = 2 or 3) and (PE5 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PE6 (PROXY = 2 or 3) and (PE6 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PE7 (PROXY = 2 or 3) and (PE7 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_PE8 (PROXY = 2 or 3) and (PE8 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_FV1 (PROXY = 2 or 3) and (FV1 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_FV2 (PROXY = 2 or 3) and (FV2 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_FV3 (PROXY = 2 or 3) and (FV3 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent.

- 18 -

Code Rule Type Description CC_P_FV4 (PROXY = 2 or 3) and (FV4 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_SK4 (PROXY = 2 or 3) and (SK4 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_AL1 (PROXY = 2 or 3) and (AL1 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_AL2 (PROXY = 2 or 3) and (AL2 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_AL3 (PROXY = 2 or 3) and (AL3 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_AL4 (PROXY = 2 or 3) and (AL4 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_AL5 (PROXY = 2 or 3) and (AL5 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_AL6 (PROXY = 2 or 3) and (AL6 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_SS1 (PROXY = 2 or 3) and (SS1 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_SS2 (PROXY = 2 or 3) and (SS2 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_SS3 (PROXY = 2 or 3) and (SS3 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_IC1 (PROXY = 2 or 3) and (IC1 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_IC2 (PROXY = 2 or 3) and (IC2 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent. CC_P_IC3 (PROXY = 2 or 3) and (IC3 ≠ -1) W The variable is not recommended to be filled in by a proxy respondent.

Code Rule (indication of problem)

Type Description of the rule Note

CC_HH01N {none of HHNBPERS_0_4, HHNBPERS_5_13, HHNBPERS_14_15, HHNBPERS_16_24, HHNBPERS_25_64, HHNBPERS_65plus is missing/’-1’} and (HHNBPERS ≠ HHNBPERS_TN)

E The total number of people living in the household is not equal to sum of number of people by age categories.

Where: HHNBPERS_TN = (HHNBPERS_0_4 + HHNBPERS_5_13 + HHNBPERS_14_15 + HHNBPERS_16_24 + HHNBPERS_25_64 + HHNBPERS_65plus). Remark: the respondent should be counted in HHNBPERS and also in one of the age subcategories.

- 19 -

CC_HH01M HHNBPERS ≠ -1 and{at least one of HHNBPERS_0_4, HHNBPERS_5_13, HHNBPERS_14_15, HHNBPERS_16_24, HHNBPERS_25_64, HHNBPERS_65plus is missing/’-1’} and (HHNBPERS < HHNBPERS_TM )

E The total number of people living in the household is not equal to sum of number of people by age categories.

Where: HHNBPERS_TM = (HHNBPERS_0_4 + HHNBPERS_5_13 + HHNBPERS_14_15 + HHNBPERS_16_24 + HHNBPERS_25_64 + HHNBPERS_65plus) and missing variables in the sum are treated as zero.

CC_HH02N {none of HHNBPERS_16_24, HHNBPERS_25_64, HH_ACT, HH_INACT is missing/’-1’} and (HH_TAN ≠ 0)

E The total number of people aged 16-64 is not equal to sum of economically active and inactive people in the same age category.

Where HH_TAN = (HHNBPERS_16_24 + HHNBPERS_25_64 - HH_ACT - HH_INACT)

CC_HH10 {(HHNBPERS = 1) and (HHTYPE ≠ 10)} or {(HHNBPERS ≠ 1) and (HHTYPE = 10)}

E One-person household not properly coded.

- 20 -

CC_HH21 (HHTYPE = 21) and {(HHNBPERS = 1) or (HHNBPERS_14_15 = 0 and HHNBPERS_16_24 = 0 and HHNBPERS_25_64 = 0 and HHNBPERS_65plus = 0) or (HHNBPERS_0_4 = 0 and HHNBPERS_5_13 = 0 and HHNBPERS_14_15 = 0 and HHNBPERS_16_24 = 0)}

E Inconsistency between values in HHTYPE and HHNBPERS.

Assumption that lone parent/couple family can be formed by persons from the age of 14 years.

CC_HH22 (HHTYPE = 22) and {(HHNBPERS = 1) or (HHNBPERS_14_15 = 0 and HHNBPERS_16_24 = 0 and HHNBPERS_25_64 = 0 and HHNBPERS_65plus = 0) or [(HHNBPERS_0_4 >= 1 and HHNBPERS_0_4 <= 98) or (HHNBPERS_5_13 >= 1 and HHNBPERS_5_13 <= 98)]}


- 21 -

CC_HH23 (HHTYPE = 23) and {(HHNBPERS = 1 or HHNBPERS = 2) or (HHNBPERS_14_15 = 0 and HHNBPERS_16_24 = 0 and HHNBPERS_25_64 = 0 and HHNBPERS_65plus = 0) or (HHNBPERS_0_4 = 0 and HHNBPERS_5_13 = 0 and HHNBPERS_14_15 = 0 and HHNBPERS_16_24 = 0)}


CC_HH24 (HHTYPE = 24) and {(HHNBPERS = 1 or HHNBPERS = 2) or (HHNBPERS_14_15 = 0 and HHNBPERS_16_24 = 0 and HHNBPERS_25_64 = 0 and HHNBPERS_65plus = 0) or (HHNBPERS_0_4 = 0 and HHNBPERS_5_13 = 0 and HHNBPERS_14_15 = 0 and HHNBPERS_16_24 = 0)}


- 22 -

CC_HHMD (HHNBPERS = 1) and (MARSTADEFACTO = 1)

E 1-person household but with a consensual union at the same time.

CC_HS_MAINSTAT

(MAINSTAT = 33) and (HS2 = 2 or HS3 = 3)

W Disabled people without activity limitation or long-term health problems.

CC_HS2_CD (HS2 = 2) and {at least one of CD1A to CD1O = 1}

W At least one of the chronic conditions reported but not a long-standing health problem in general.

CC_HS3_PL (HS3 = 3) and (PL1 = 3 or PL2 = 4 or PL3 = 3 or PL4 = 4 or PL5 = 4 or PL6 = 4 or PL7 = 4)

W Severe problems with physical or sensory functions but not activity limitations in general.

CC_HS3_PC (HS3 = 3) and (PC1A = 4 or PC1B = 4 or PC1C = 4 or PC1D = 4 or PC1E = 4)

W Severe problems with personal care activities but not activity limitations in general.

CC_HS3_HA (HS3 = 3) and (HA1A = 4 or HA1B = 4 or HA1C = 4 or HA1D = 4 or HA1E = 4 or HA1G = 4)

W Severe problems with household activities but not activity limitations in general.

‘Difficulty in doing occasional heavy housework’ is not part of the rule.

CC_AC_HC (AC2 = 1 or 2) and {HO1 ≠ 1 and HO3 ≠ 1 and (AM1 ≠ 1 or 2) and AM2 ≠ 1 and AM4 ≠ 1 and AM6A ≠ 1 and AM6B ≠ 1 and AM7 ≠ 1}

W A treatment after an injury reported but use of any health care not reported in the past 12 months.

The records marked with the warning are recommended to be checked all.

CC_PL4_PL5 (PL4 > PL5) and (PL4 = 1 or 2 or 3 or 4) and (PL5 = 1 or 2 or 3 or 4)

W PL4 should give lower values than PL5 (less severe conditions for evaluation of hearing)


- 23 -

CC_PA1A {none of PA1_YYYYMM or REF_YYYYMM is blank} and (PA1_YYYYMM > REF_YYYYMM)

E The date of vaccination is not compatible with the date of the interview.

Where: - PA1_YYYYMM is PA1 transformed into YYYYMM (if PA1 is in MMYYYY format and both MM and YYYY have valid values – valid values for MM are 01 – 12 and for YYYY are 2010 – 2015) or blank (if PA1 is -1 or 0); - REF_YYYYMM is a combination of REFYEAR and REFMONTH transformed into YYYYMM (leading zeros are important for deriving the variables) (if none of REFYEAR and REFMONTH is -1) or blank (if REFYEAR or REFMONTH are -1).

CC_PA1B {none of PA1_YYYY or REF_YYYY is blank} and (PA1_YYYY > REF_YYYY)

E The year of vaccination is not compatible with the year of the interview.

Where: - PA1_YYYY is PA1 transformed into YYYY (if PA1 is in MMYYYY format) or blank (if PA1 is -1, 0, or MM does not contain valid values – valid values for MM are 01 – 12 – or is missing); - REF_YYYY is REFYEAR (if REFYEAR is not -1) or blank (if REFYEAR is -1).

CC_PA2 (PA2 = 1) and {HO1 ≠ 1 and HO3 ≠ 1 and AM2 ≠ 1 and AM4 ≠ 1 and AM7 ≠ 1}

W A blood pressure measurement reported but no relevant health care treatment (out of those services included in the survey) in the past 12 months.


W A blood cholesterol measurement reported but no relevant health care treatment (out of those services included in the survey) in the past 12 months.

- 24 -


W A blood sugar measurement reported but no relevant health care treatment in (out of those services included in the survey) the past 12 months.

CC_PA5 (PA5 = 1) and {HO1 ≠ 1 and HO3 ≠ 1 and AM2 ≠ 1 and AM4 ≠ 1}

W A faecal occult blood test reported but no relevant health care treatment (out of those services included in the survey) in the past 12 months.


W A colonoscopy reported but no relevant health care treatment (out of those services included in the survey) in the past 12 months.


W A mammography reported but no relevant health care treatment (out of those services included in the survey) in the past 12 months.


W A cervical smear test reported but no relevant health care treatment (out of those services included in the survey) in the past 12 months.

- 25 -

CC_BM {none of BM1 or BM2 is -1} and (BMI < 14) or (BMI > 45)

W Possible extreme combination of height and weight

Where BMI is calculated as follows: IF {none of BM1 or BM2 is -1} THEN BMI = [BM2 / (BM1 * BM1/10000)] ELSE: BMI = blank The records marked with the warning are recommended to be checked all.

CC_AL1_1 (AL1 = 1) and (AL24T < 6) and (AL2 ≠ -1) and (AL4 ≠ -1)

W Numbers of days of alcohol consumption do not correspond between AL1 and AL2/AL4.

Where: - AL2T = 5 - AL2 (if 1<=AL2<=5) or AL2T = 0 (otherwise). - AL4T = 4 – AL4 (if 1<=AL4<=4) or AL4T = 0 (otherwise). - AL24T = AL2T + AL4T. Not (necessarily) '7' as AL1 category 1 refers to 'every day or almost every day'. Skip the check if AL2 or AL4 are missing. The records marked with the warning are recommended to be checked all.

CC_AL1_2 (AL1 = 2) and {(AL24T < 5) or (AL24T > 6)} and (AL2 ≠ -1) and (AL4 ≠ -1)


Skip the check if AL2 or AL4 are missing. The records marked with the warning are recommended to be checked all.







- 26 -

CC_AL6 {AL1 = 1, 2, 3, 4, 5, 6 or 7 and AL6 = 1, 2, 3, 4, 5, 6 or 7} and (AL6 < AL1)

W The number of days of risky alcohol consumption is higher than the usual number of days of alcohol consumption.


CC_AL23 (AL2 = 1 or 2 or 3 or 4) and (AL3 = 8)

W Drinking from Monday to Thursday but 0 drink a day.


CC_AL45 (AL4 = 1 or 2 or 3) and (AL5 = 8)

W Drinking from Friday to Sunday but 0 drink a day.


CC_PE1_W (PE6 = 1) and (PE7M > 480)

W The time spent doing sport is more than 8 hours per day.

Where PE7M is PE7 transformed into minutes: HH(part) * 60 + MM(part). The records marked with the warning are recommended to be checked all.

CC_PE2_W (PE6 = 2) and (PE7M > 960)



CC_PE3_W (PE6 = 3) and (PE7M > 1440)



CC_PE4_W (PE6 = 4) and (PE7M > 1920)



CC_PE5_W (PE6 = 5) and (PE7M > 2400)



CC_PE6_W (PE6 = 6) and (PE7M > 2880)



CC_PE7_W (PE6 = 7) and (PE7M > 3360)



CC_PE1_E (PE6 = 1) and (PE7M > 960)

E The time spent doing sport is more than 16 hours per days.

CC_PE2_E (PE6 = 2) and (PE7M > 1920)

E The time spent doing sport is more than 16 hours per day.

CC_PE3_E (PE6 = 3) and (PE7M > 2880)


- 27 -

CC_PE4_E (PE6 = 4) and (PE7M > 3840)


CC_PE5_E (PE6 = 5) and (PE7M > 4800)


CC_PE6_E (PE6 = 6) and (PE7M > 5760)


CC_PE7_E (PE6 = 7) and (PE7M > 6720)


CC_PE1_PL (PE1 = 3) and (PL6 = 4 or PL7 = 4)

W Severe problems with physical or sensory functions but performing mostly heavy labour or physically demanding work.

CC_PE1_PC (PE1 = 3) and (PC1A = 4 or PC1B = 4 or PC1C = 4 or PC1D = 4 or PC1E = 4)

W Severe problems with personal care activities but performing mostly heavy labour or physically demanding work.

- 28 -

Countries which were granted derogations4 use the same microdata structure as other countries. For variables for which derogations were granted, the code ‘-1’ shall be used. All rules using the variables under derogations should not be applied to these countries.

The countries and variables for which derogations were granted are listed here:

Country Derogation

Belgium

— INTLANG (interview language);

— PL4, PL5, PL7 of the set on physical and sensory functional limitations;

— MH1a till MH1h on mental health;

— AM3 and AM5 on ambulatory care;

— PA4 on preventive activities;

— UN1a till UN2d on unmet needs.

France

— AL1 till AL6 on alcohol consumption;

— IC1 till IC3 on informal care.

the Netherlands

— PL5 and PL6 on physical and sensory functional limitations;

— MH1a till MH1h on mental health;

— PE1 till PE8 on physical activities;

— AL1 and AL6 on alcohol consumption.

4 Commission Implementing Decision of 19 February 2013 granting derogations to certain Member States with respect to the transmission of statistics pursuant to Regulation

(EC) No 1338/2008 of the European Parliament and of the Council on Community statistics on public health and health and safety at work, as regards statistics based on the European Health Interview Survey (EHIS).



EUROPEAN HEALTH INTERVIEW URVEY - circabc.europa.eu wave 2... · • SC = Skip checks, which follows the filters defined in the annex of the EHIS Regulation; • CC = Consistency

Documents