The effect of postal questionnaire burden on response rate and answer patterns following admission to intensive care: a randomised controlled trial

Robert Hatch; Duncan Young; Vicki Barber; David A Harrison; Peter Watkinson

doi:10.1186/s12874-017-0319-3

The effect of postal questionnaire burden on response rate and answer patterns following admission to intensive care: a randomised controlled trial

BMC Med Res Methodol. 2017 Mar 27;17(1):49. doi: 10.1186/s12874-017-0319-3.

Authors

Robert Hatch^{1

2}, Duncan Young^{1

3}, Vicki Barber^{1

4}, David A Harrison^{1

5}, Peter Watkinson^{6

7}

Affiliations

¹ Kadoorie Centre for Critical Care Research and Education, University of Oxford, John Radcliffe Hospital, Headley Way, Oxford, OX3 9DU, UK.
² Registrar in Anaesthetics, Oxford Deanery, Oxford, UK.
³ Intensive Care Medicine, University of Oxford, Oxford, UK.
⁴ OCTRU Hub, University of Oxford, Oxford, UK.
⁵ Statistician, Intensive Care National Audit and Research Centre (ICNARC), London, UK.
⁶ Kadoorie Centre for Critical Care Research and Education, University of Oxford, John Radcliffe Hospital, Headley Way, Oxford, OX3 9DU, UK. peter.watkinson@ndcn.ox.ac.uk.
⁷ University of Oxford and Consultant Intensive Care Physician, Oxford University Hospitals NHS Trust, Oxford, UK. peter.watkinson@ndcn.ox.ac.uk.

Abstract

Background: The effects of postal questionnaire burden on return rates and answers given are unclear following treatment on an intensive care unit (ICU). We aimed to establish the effects of different postal questionnaire burdens on return rates and answers given.

Methods: Design: A parallel group randomised controlled trial. We assigned patients by computer-based randomisation to one of two questionnaire packs (Group A and Group B).

Setting: Patients from 26 ICUs in the United Kingdom.

Inclusion criteria: Patients who had received at least 24 h of level 3 care and were 16 years of age or older. Patients did not know that there were different questionnaire burdens. The study included 18,490 patients. 12,170 were eligible to be sent a questionnaire pack at 3 months. We sent 12,105 questionnaires (6112 to group A and 5993 to group B).

Interventions: The Group A pack contained demographic and EuroQol group 5 Dimensions 3 level (EQ-5D-3 L) questionnaires, making four questionnaire pages. The Group B pack also contained the Hospital Anxiety and Depression Score (HADS) and the Post-Traumatic Stress Disorder Check List-Civilian (PCL-C) questionnaires, making eight questionnaire pages in total.

Main outcome measure: Questionnaire return rate 3 months after ICU discharge by group.

Results: In group A, 2466/6112 (40.3%) participants responded at 3 months. In group B 2315/ 5993 (38.6%) participants responded (difference 1.7% CI for difference 0-3.5% p = 0.053). Group A reported better functionality than group B in the EQ-5D-3 L mobility (41% versus 37% reporting no problems p = 0.003) and anxiety/depression (59% versus 55% reporting no problems p = 0.017) domains.

Conclusions: In survivors of intensive care, questionnaire burden had no effect on return rates. However, questionnaire burden affected answers to the same questionnaire (EQ-5D-3 L).

Trial registration: ISRCTN69112866 (assigned 02/05/2006).

Keywords: Critical care; Critical illness; EQ5D; Health related quality of life; Intensive care; Multicentre study; Outcome assessment; Questionnaire; Randomised control trial; Response rate; Survivors.

Publication types

Multicenter Study
Randomized Controlled Trial

MeSH terms

Aged
Critical Care
Critical Illness / therapy*
Female
Hospitalization
Humans
Male
Middle Aged
Outcome Assessment, Health Care / methods*
Outcome Assessment, Health Care / statistics & numerical data*
Patient Compliance / statistics & numerical data*
Patient Discharge
Quality of Life
Surveys and Questionnaires*

Associated data

ISRCTN/ISRCTN69112866