5 Our traditional approach Standard interviewer visit + some measures (height and weight) Nurse follow up visit at participant’s home to collect other measures Interviewer introduces nurse visit, may make appointment NatCen first used nurses in 1994 Health Survey for England Requires a special panel of freelance nurses (we have around 150)
6 Drawbacks of nurse follow ups: drop out HSE 2010 (adults) 71% (5,587 visits) % of interviewed who have nurse visit Survey NDNS Yr 3 74% (698 visits) ELSA wave 4 86% (8,218 visits) US (W2Y2 + W3Y1) 65% (4,409 visits) Nurse HSE sample very similar to interview: Slightly fewer smokers and non- drinkers (esp. men). Little difference by Age; sex; BMI; general health; diet. Slightly more: Men with high BP
7 We have used interviewers to collect… Height Weight Waist Saliva sample Urine samples Timed walks Infant length & head circumference Grip strength Cognitive tests And recently developed protocols to extend interviewer biomarker collection to: Blood pressure (Scottish Health Survey) Dried blood spots (Scottish Health Survey from 2013)
8 Drawbacks of nurse follow ups: logistics Child fasting blood (35%) NDNS requires clinics within two hours of each sample point Specialist phlebotomists for some visits Blood to lab within 2 hours for spinning Nurse visit (74%) Adult fasting blood (69%) Child 24 hr urine (74%) Adult 24 hr urine (80%) NDNS Interviewer visit Paediatric phlebotomist
9 Use of clinics Diet and Nutrition Survey of Infants and Young Children Blood taking from 4 – 18 month olds, immediate spinning of blood Feed special tracer water to these infants, specific dose Skinfold measures Decision to use clinic-based follow-up: 20 clinics recruited (hospital/private), specific training for staff 2 mobile vans with experienced staff (initially for people living far from a clinic)
Does it matter who takes the measurements or how they take it? 3.
12 …evidence from the Scottish Health Survey Large scale survey monitoring the health of the Scottish population Carried out 1995, 1998, 2003 and annually since 2008 Collecting biomeasures using the traditional nurse model since 1995 Switched to using field interviewers in 2012 What impact, if any, will this move have on our ability to continue to monitor trends in…
16 …blood pressure Protocol unchanged Concerns about non-medically trained interviewers collecting it Interest in whether “white coat syndrome” extends to nurses
17 Validation study objectives were… To establish the degree of correlation between interview and nurse measurements Generate calibration equations that allow for SHeS statistics to be ‘converted’
18 How did we do it? Paired up trained field interviewers and survey nurses (15 pairs) Pairs visited respondent in own home, taking turns to take BP readings & waist measurements Purposive sample – oversampled aged 55+ 300 paired interviews carried out (Mar-May 2012)
19 Factors considered Order of interviews was randomised Carried out consecutively Partner not in the home at time of interview Same blood pressure machine used Participant told whether an interviewer/nurse Results from 1st visit not given to participant
So did it matter who took the measurements & how?
22 Biomarkers – things to consider It is possible to collect a range of biological measures within a survey based study, but it is important to consider how: Who will collect (interviewer, nurse, clinic, self): Complexity of measure, protocol, acceptability, analysis requirements, target population Ethical considerations Training is vital – quality is important! Equipment: ease of use, portability, durability Labs – work with them to agree protocols, check quality! Data analysis – can be a minefield.
If you want further information or would like to contact us Lisa Rutherford Research Director T. +44 (0)131 221 2555 E. firstname.lastname@example.org Visit us online, natcen.ac.uk Thank you