Patient‐reported symptoms and experience following Guillain‐Barré syndrome and related conditions: Questionnaire development and validation

Abstract Background Guillain‐Barré syndrome (GBS) is a rare inflammatory peripheral nerve disorder with variable recovery. Evidence is lacking on experiences of people with GBS and measurement of these experiences. Objective We aimed to develop and validate an instrument to measure experiences of people with GBS. Design We used a cross‐sectional design and online self‐administered questionnaire survey. Question domains, based on a previous systematic review and qualitative study, covered experiences of GBS, symptom severity at each stage, healthcare and factors supporting or hindering recovery. Descriptive, exploratory factor and reliability analyses and multivariable regression analysis were used to investigate the relationships between variables of interest, explore questionnaire reliability and validity and identify factors predicting recovery. Setting and Participants People with a previous diagnosis of GBS were recruited through a social media advert. Results A total of 291 responders, of different sexes, and marital statuses, were included, with most diagnosed between 2015 and 2019. Factor analysis showed four scales: symptoms, information provided, factors affecting recovery and care received. Positive social interactions, physical activity including physiotherapy and movement, changes made at home and immunoglobulin treatment were important for recovery. Multivariable models showed that immunoglobulin and/or plasma exchange were significant predictors of recovery. Employment and recovery factors (positive interactions, work support and changes at work or home, physical activity and therapy), though associated with recovery, did not reach statistical significance. Conclusion The questionnaire demonstrated good internal reliability of scales and subscales and construct validity for people following GBS. Patient Contribution Patients were involved in developing and piloting the questionnaire.

with recovery demonstrated the importance of early diagnosis, positive experiences of inpatient care, active support for recovery and good communication and information provision. 13 Quality in healthcare is widely considered to consist of three interrelated components: safety, effectiveness and experience. 14,15 Tools such as the Inflammatory Rasch-built Overall Disability Scale, the Medical Research Council sum score and the Inflammatory Neuropathy Cause and Treatment disability score have been used to monitor the effectiveness of treatment and disease progression. 16 Healthcare experiences include 'experiences of what health services and staff are like and do' and experiences of how they feel services 'enable [them] to be and do what [they] value being and doing within and beyond [their] healthcare encounters'. 17 Patient-reported experience measures (PREMs) are widely used to assess patient experience as a key aspect of quality. 18,19 This is also relevant and important for patients' experiences of conditions such as GBS and its variants.
We aimed to develop and validate a questionnaire to quantify experiences of people with GBS.

| Design
We used a cross-sectional design using a self-administered online questionnaire survey designed to explore symptoms, care experiences and recovery in people who previously had GBS.

| Questionnaire development
Questionnaire domains and items were based on a systematic review and metasynthesis of qualitative studies 11 and an interview study of people with the condition. 13 The domains for people who previously had GBS included participant characteristics; severity of symptoms (physical, psychological and social) at each stage of illness; medical health-seeking experience; treatment and care experiences; followup and support; and social or work-related experience.
The initial questionnaire was piloted with four people who had recovered from GBS, of whom two had taken part in an earlier interview study. 13 The questionnaire was also discussed with the Guillain-Barré Syndrome and Associated Inflammatory Neuropathies (GAIN) charity, the Healthier Ageing Patient and Public Involvement group at the University of Lincoln and members of the research team.
Comments and suggestions were used to revise some of the questions to ensure that they were appropriate for the intended population of GBS patients.

| Data analysis
The internal consistency of the GBS questionnaire was assessed using Cronbach's α. 20 This test was used to establish the level of agreement between items belonging to the same scale. Four main scales were developed, which contained items scored on a 7-point Likert scale including symptoms, care received, factors affecting recovery and information provided. Some of these scales were divided into further subscales: initial, in-hospital, after-hospital and current symptoms as well as care received in hospital and after discharge from hospital.
Factor analyses were run to identify questionnaire subscales.
The scales included were suitable for this type of analysis as indicated by the Kaiser-Meyer-Olkin measure, which was higher than 0.7 for all of them. Retained factors were those with eigenvalues greater than 1 and items with loadings higher than 0.4. 21 As such, the scales included in the factor analysis (FA) were symptoms (initial, in hospital, residual and current), factors affecting recovery and information provided.
Multivariate linear regression models were used to identify the factors predicting recovery. Two regression models were run: The first one using the scales of the questionnaire as predictors and the second one using the subscales derived from factor analyses as the main predictors together with demographic characteristics that might have influenced the outcome. These demographic predictors included age and the binary variables: sex (female or male), employment status (employed or unemployed) and living with someone else or alone. The recovery score, which was used as the main outcome, was computed using the formula: recovery score = mean score of in hospital symptoms − mean score of present symptoms.
The assumption of normality was met as indicated by both histograms and P-P plots of residuals. Homoscedasticity was present as indicated by scatterplots. The assumption of no multicollinearity was also met for both models as indicated by Durbin Watson tests with values close to 2 (1.93 for the first model and 2.12 for the second model), tolerance values higher than 1 and Variance Inflation Factor values smaller than 10.

| Responder characteristics
In total, 291 participants responded fully or partially to the questionnaire. Table 1

| Reliability of scales and subscales
The reliability (internal consistency) of the main scales was excellent for symptoms (α > .9), good for information provided and for factors affecting recovery (α > .8) and acceptable for care received (α ≥ .7).
Importantly, none of the scales had poor reliability (α ≤ .6). 20 Overall, these results presented in Table 3 indicate that the questionnaire was a reliable measure, with good internal consistency.  Table S1.
The internal reliability of the new subscales is presented in Table 4.
The reported severity of symptoms for each subscale at different time points (initial before admission to hospital, in hospital, residual and current, i.e., when responders were completing the questionnaire) indicated that symptoms were most severe when responders were in hospital and those affecting the peripheral nervous system were most prominent (Figure 1).
Responders were more satisfied with the information provided by specialists rather than nonspecialists (Figure 2 and Table S2).
A combination of physical, psychological and social factors was associated with recovery; these factors were identified following the FA and an average score was calculated for each factor; details can be seen in Table S3. The factors considered by responders to be most important for recovery were positive social interactions, physical activity including physiotherapy and movement, changes made at home and immunoglobulin treatment (Figure 3).

| Prediction models
Multivariable regression models were fitted to the data. The predictors used were gender, age, employment status and living alone or with someone, since these variables had been previously shown to influence recovery. 11,13 In the first regression model, the main scales of the questionnaire (shown in Table 5) were included as predictors. In the second regression model, the main subscales of the questionnaire (shown in Table 6) were included in the model. The results indicated that immunoglobulin and/or PE treatment were significant predictors of recovery. Factors considered by responders to be most important for recovery were positive social interactions, physical activity, changes made at home and immunoglobulin treatment. Responders were more satisfied with information provided by specialists rather than nonspecialists. Multivariable models showed that immunoglobulin and/or PE treatment were significant predictors of recovery. 4 Being in employment and recovery factors in combination (positive social interactions, support and changes at work support, changes at home, physical activity and counselling or occupation therapy) were positively associated with recovery, but this did not reach statistical significance.

| Comparison with the existing literature
Although many people with GBS are told that they will recover and some do so completely, many are still affected in the longer term.
Early results from the largest ongoing prospective study, the IGOS, 7 have shown that 8% could not walk and 7% had died at 1 year, with wide international variations in outcome. 8 Previous studies have also shown long-term neurological deficits in most patients after a year or beyond. 22,23 Furthermore, a third had changed work or were affected in their functional ability and half had altered their leisure activities. 22 Psychological 24 and social dysfunction 25 often persist longer term, affecting health-related quality of life. 26 Previous research has suggested a wide variation in positive and negative experiences at various stages of treatment and recovery from GBS. 11,13 We also found wide variations in experiences of care from different healthcare professionals during the illness journey, with the most positive experiences of care in hospital, from consultants, followed by nurses and therapists. Consultants, followed by physiotherapists were also rated highly for care at follow-up, and although in this study physical, psychological and social support were (nonsignificantly) associated with improvement in symptoms, experiences of care and psychosocial support remain important aspects of quality of care.
Rehabilitation studies, involving careful follow-up, show positive benefits of rehabilitation on function 27 and mortality 28 before discharge from hospital, but intensive physiotherapy beyond 6 months was also found to improve functional outcomes. 6 Responders in our study valued physiotherapy and perceived this to improve their recovery, but shortfalls in provision for both inpatient and outpatient rehabilitation have been found in previous studies. 29,30 Positive social interactions and changes at home were also associated with recovery in this study. Positive social interactions include family or peer support. 13 A systematic review found that peer support as a potential intervention for recovery in critical care populations reduced psychologic morbidity and improved self-efficacy, although the quality of included studies was low. 31 Finally, complementary therapies such as acupuncture, vitamins and hyperbaric oxygen have been used as an adjunct to conventional treatment, but the only nonrandomized study was deemed of low quality. 32

| Strengths and limitations
The number of questionnaires returned was sufficient for the planned analysis, and most participants who began completed the ques- with recovery. 11,13 Because the survey showed good evidence of face and construct validity and internal consistency, it could be used to assess patient experience and how experience of care and support could be improved in a larger population of people with GBS.
Further research needs to be done to develop patient-reported outcome measures 33 and PREMs for GBS beyond traditional disability measures such as the GBS Disability Scale. 34 The experience scales developed in this survey could be used to develop and evaluate the effect of interventions designed to improve experiences at various stages of treatment and recovery including in the longer term, including better access to rehabilitation and innovative social interventions such as peer or employer support.

| CONCLUSION
Our findings showed that the GBS patient experience survey showed characteristics of a good measure, with evidence of internal consistency and construct validity. The GBS patient experience questionnaire should be tested more widely to seek further evidence of reliability, construct validity and sensitivity to differences in care and setting.