Keywords: Sickness Impact Profile; Questionnaires; Validation; Reliability

Introduction

Health-related quality of life (HRQOL) is important in a variety of diseases and one of the most important outcome measures after an intervention. Generic questionnaires are important as they measure all the aspects of the disease and reflect the overall impact of the disease and the benefits of treatment [1].

The Sickness Impact Profile (SIP) is a generic questionnaire of health-related functional status [2] and can be used across different types and severities of diseases [3]. SIP scores are available for approximately 18 different diseases or populations [4]. In order to measure the health status, participants must be questioned in their own language [5].

SIP has been translated into several languages including Arabic, Chinese for Hong-Kong, Danish, Dutch, Dutch for Belgium, English for Mexico, English for the UK, Finnish, French, French for Belgium, Italian, German, Norwegian, Portuguese, Romanian, Russian, Spanish, Spanish for Mexico, Spanish for the USA, Tamil and Thai, Swedish [6]. The aim of this study was the cross-cultural adaptation of SIP in Greek language (SIP-GR) and the assessment of its psychometric properties.

Materials and Methods

The study was approved by the Cyprus Bioethics Committee (EEBK EΠ 2018.01.148) and whole adaptation process was approved by the SIP developers.

Sample size calculation

Sample size calculations were based on the intraclass correlation coefficient (ICC) of test-retest reliability of total scores. Using an acceptable ICC of at least 0.7, an expected ICC of at least 0.9 a power of 80% a significance level of 0.05 the sample size required is 19 subjects [7]. An additional twenty percent of drop-out rate was included bringing the total number of subjects required to 21. We recruited ninety subjects all together, sixty patients with cardiopulmonary and musculoskeletal diseases and thirty apparently healthy but obese or overweight subjects (at risk group).

Participants

Ninety (n=90) participants were included in this study after signing a written informed consent. Recruitment was achieved via advertisement in the local area of Nicosia, Cyprus and from the patient list of a local Physiotherapy clinic. Participants were invited to participate in this study if they met the following inclusion criteria: ≥18 years old, comprehension of Greek language, obese and overweight (≥25 body mass index), had recently undergone cardiac surgery (within 2-months after surgery), suffer from primary osteoarthritis of the hip or knee joint with symptoms lasting at least 2-months, complained for nonspecific low back pain lasting at least 2-months or were diagnosed with pulmonary diseases with recurrent exacerbations. All the participants (n=60) except overweight and obese, attended 2 weekly conservative physiotherapy sessions according to their diagnosis, during the period of data collection. Participants reported information on sociodemographic variables such as gender, age and occupation. In order to examine the test-retest reliability, all subjects were asked to fill the questionnaire twice at one-week interval. In order to assess its validity, SIP was compared to 36-item Short Form Health Survey (SF-36).

Cross-cultural adaptation process

In general, cultural adaptation of SIP followed the internationally accepted guidelines for cultural adaptation of patient reported outcomes published by Beaton et al. [8]. The process intended to produce equivalency between the original and translated versions in terms of content and to adapt it culturally so that the original meaning and intent of the items are maintained. The process involved the following 6 steps:

Step 1: Forward translation and harmonization of forward translation

Two of the authors, both bilingual physiotherapists, native in Greek language independently translated the original version of SIP into Greek and produced a report. The two versions were then synthesized into one initial translation by consensus of the two reviewers and a new report was produced which was send to the SIP developers.

Step 2: Backward translation and harmonization of backward translation

Two bilingual translators one physiotherapist and one language expert, native in English language and fluent in Greek, produced two independent back translations of the original Greek version. All translators were kept blind to the original English version of SIP.

An expert committee consisting of the translators, one more Physiotherapist and one medical doctor produced the prefinal version of the SIP-GR. The committee made a great effort to ensure semantic, idiomatic, experimental and conceptual equivalence existed between the original and the SIP-GR versions. The whole process was documented and a report was again sent to SIP developer.

Step 3: Validation of the pre final version

Validation of the translation was performed by evaluating the comparability of language and the similarity of interpretation using a 7-point Likert scale ranging from 1 (extremely comparable/ similar) to 7 (not at all comparable/not at all similar). 10 bilingual individuals independently compared the English and the translated versions item by item and rated each one in terms of comparability and similarity. Any item above 3 in comparability or 2.5 in similarity was deemed appropriate for revision. No item needed revision. The same 10 individuals were used for cognitive debriefing. Cognitive equivalence of the translated version was tested between various educational backgrounds and different regions of Greece in order to capture differences in dialects among individuals.

Step 4: Expert committee review

The committee assessed all the reports from the previous steps and made all the necessary modifications to optimize the final version.

Step 5: Proofreading

A proofreading company was consulted to correct the final version in terms of spelling, diacritical, grammatical, or any other errors. Following this step, the Final version was ready for pilot testing.

Step 6: Pretesting (pilot study)

The pilot testing of the questionnaire was performed in 10 individuals (7 males, 3 females) from the general population approached randomly in a local mall. Individuals had to be overweight or obese and be able to speak, read and understand Greek. The 10 participants (mean age 47.7 ± 22.64 years) completed the self-administrated SIP-136 Greek version of the questionnaire. Eight participants were obese, 2 were overweight, 7 had hypertension, but only 5 of them were taking medication, 4 participants had osteoarthritis (OA), 1 was a stroke survivor, 1 was diagnosed with Parkinson’s disease (PD), 1 with asthma and 1 had a stent due to a congenital heart disease. All individuals were asked to complete the questionnaire without the help of an interviewer and record any difficulties in comprehending any item on a standard form. SIP scores and the time needed to fill the questionnaire were also recorded. None of the participants recorded any problem with comprehending any item and none left any item unanswered. The mean time to complete the questionnaire was 34.2 ± 7.32 minutes and the mean total score was 14.8 ± 9.37 (Physical component = 14.6 ± 16.43, Psychosocial component = 12.5 ± 7.88).

Measurement instruments

Sickness Impact Profile

SIP is a generic questionnaire designed to subjectively assess the physical and psychological functioning in a wide range of diseases. SIP consists of 136-items that are divided into 12 categories related to daily living which are then grouped into physical and psychosocial dimensions [9,10]. Physical domain includes: ambulation, mobility, body care/movement. Psychosocial domain includes: social interaction, communication, alertness behaviour, emotional behaviour. Sleep and rest, eating, home management, recreation and pastimes, and work are considered independent categories. Each item is a question in present tense and patients is asked to reply how they feel at the time of test administration. Replies in all questions are in binary form (΄΄Yes/ No’’) and the patient selects all the questions that are applicable to them. The total score is calculated by summing up the domain scores and the result is expressed as a percentage of the maximum possible score based on the answers (0-100). A higher score represents a more severe impact of the disease on health.

36-item Short Form Health Survey

The 36-item Short Form Health Survey SF-36 (SF-36) is a generic questionnaire which includes 36 questions based on the general health status, divided into 8 subscales which ultimately provide two scores of physical and mental health status [11,12]. SF-36 subscale scores, range between 0 and 100, where greater score shows better HRQOL [13].

Assessment Analysis

Reliability Test-retest reliability

The two-way mixed model, intraclass correlation coefficient (ICC), with absolute agreement was used to assess the reliability between the first and the second time point in the domain and overall scores of SIP-GR (1-week interval). If ICC values were ≤ 0.40 reliability was considered poor, between 0.40-0.75 moderate, between 0.75 and 0.90 substantial and > 0.90 excellent [14]. Because most subjects received treatment during the one-week test/retest period, all participants were asked to subjectively rate their change of health status during the last week using a 0-100% scale. Test-retest reliability was calculated separately on those subjects that rated their health status change as zero (n=36). Those patients were all the obese (no treatment), 5 musculoskeletal and 1 Cardiopulmonary patient. Moreover, the standard error of measurement (SEM) was calculated based on the formula SEM=SD∗√(1−ICC), where SD is the standard deviation of the initial assessment, and ICC is the value obtained by the analysis of test-retest. In addition, minimal detectable change at the 90% confidence level (MDC) was calculated based on the MDC= 1,65 x √2 x SEM formula [15]. Lower SEM values indicates better reliability of the measure, whereas lower MDC values indicates a more sensitive measure [16].

Internal consistency

Internal consistency was determined using the Cronbach’s alpha. Values higher than 0.7 are considered as sufficient [14].

Validity

Construct validity

Spearman’s rank correlation coefficient (rs) was used to test the construct validity between SIP-GR, and SF-36 questionnaire, since the results were not normally distributed. Rs values of 0.000.30 are considered weak, 0.31-0.59 moderate and 0.60-1.00 strong [17]. The hypothesis is that there will be significant correlation between the overall scores of the two questionnaires and that the correlation of the psychological component of SIP will be higher with the Mental health of SF-36 than the Physical health and vice versa.

Known group validity

The total score and as well as the domain scores were compared between the three groups of patients namely the obese, the musculoskeletal and the cardiorespiratory group. The hypothesis is that at risk group (obese) will be significant different than the two patient groups.

Data Analysis

Data analysis was performed using SPSS (Version 25.0) and Jamovi (Version 2.2.5). The level of significance was set at 0.05. Descriptive statistics were reported using means and standard deviations (SD), or frequencies for the demographic characteristics. SIP scores are summarized for each of the 12 categories and for the domain and total score. Kruskal-Wallis test was used to assess differences between the three patient groups. Pre and post comparisons were performed to determine the changes in each questionnaire between initial assessment and re-assessment using Friedman’s Test. Test-retest reliability was assessed using ICC and 95% confidence intervals. Internal consistency was evaluated using Cronbach’s alpha. Construct validity was evaluated via Spearman’s rank correlation coefficient.

Results

A total of 90 participants included in this study. Thirty of the participants had a history of musculoskeletal problems, 30 participants had a history of cardiac or pulmonary problems and the remaining 30 participants were obese or overweight. There was a good balance between male (54.4%) and female (45.6%) gender. Table 1 summarizes the characteristics of the respondents.

Table 1: Respondents’ characteristics.

Table 2presents the means and SDs of the different categories, domains and total score of SIP for the three groups of patients at initial assessment and re-assessment. Generally, the scores were lower for obese subjects and higher for Cardiopulmonary patients. Musculoskeletal patients were lower than Cardiopulmonary group and higher than Obese group in most categories at initial assessment. However, as most of them improved with treatment they achieve lower scores than Obese in several categories, in Physical component and total score during re-assessment (Table 2).

Table 2: SIP categories, domain and total scores by patient group (Mean ± SD).

Reliability

Internal consistency

Cronbach’s alpha was calculated for the individual categories and the domain and the total scores of SIP. The overall Cronbach’s alpha for SIP-Total score >0.9, for SIP-Psychological was >0.8 and for SIP-Physical >0.9. The categories scores were >0.5 except for Communication and Work (Table 3).

Table 3: Cronbach’s alpha for categories, domain and total score of SIP.

Test-retest reliability

Intraclass correlation coefficient (ICC) was higher in the subjects that had no change in their health status compared to those that changed due to treatment. ICC for the total score was 0.691 for all subjects, 0.562 for those that reported a subjective change in their health status and 0.999 for those that reported no change in health status. This means there was moderate test-retest reliability for those that had a treatment effect and excellent reliability for those that had no change in health status. The results are similar for the Physical and Psychosocial domains and are included in (Table 4).

Table 4: SIP Test-retest reliability (ICC, 95% CI).

The effect of treatment in the questionnaire scores and the reliability of the instrument are supported by the comparisons of pre and post values among patients with significant improvement (N=54) and those without improvement (N=36). Only those with no change in health status had no significant difference between the pre and post values in both questionnaires.

Table 5: Comparison between initial assessment and re-assessment among those with health improvement and those without improvement (Median/IQR).

SIP-GR total score demonstrated an SEM of 5.5 and MDC of 12.8 points for the assessments at baseline, and an SEM of 4.6 and MDC of 10.8 points for the re-assessment using the ICC value of all subjects (0.691).

Construct Validity

There were strong negative correlations between SIP Physical component and SF36 Physical Health (rs=-0.621, p=0.001), between SIP Psychosocial component and SF36 Mental Health (rs=-0.619, p=0.001) and between SIP Total score and SF36 Total score (rs=-0.661, p=0.001) at the initial assessment (Table 6). Those correlations were still significant but only moderate (rs=-0.455, p=0.001, rs=-0.437, p=0.001, rs=-0.341, p=0.001 respectively) at reassessment (Table 7).

Table 6: Spearman’s Correlation between SIP and SF36 at initial assessment (N=90).

Table 7: Spearman’s Correlation between SIP and SF36 at re-assessment (N=90).

Participant’s self-assessment of health status

Participants were asked to grade their overall health status, post 1-week on a numeric rating scale (0-100%), where 0% indicated no degree of improvement and 100% indicated a very high degree of improvement. The percentage change (Post-Pre value/Pre Value × 100%) from baseline to endpoint for SIP and SF-36 questionnaires were calculated in 1-week interval. There were significant correlations between subjective health improvement and change in SIP and SF-36 scores as shown in (Table 8). The negative correlations between SIP and health improvement is because SIP post values were smaller that pre values, as lower SIP scores mean lower impact of disease (greater improvement).

Table 8: Spearman’s correlation between subjective health improvement and change in SIP/SF36 scores.

Known group validity

Significant differences between groups were found for domain and total scores of both SIP and SF-36. In general, impact of disease was higher for cardiopulmonary patients in both SIP Physical and Psychological component as well as total score. Impact on musculoskeletal patients was higher compared to obese except for Physical component at initial assessment and Total score at reassessment (Table 9).

The results for SF-36 were comparable to SIP. In general, obese patients showed higher functional capacity than the other two groups. Cardiopulmonary patients were significantly different from musculoskeletal patients only in terms of Physical health (Table 10).

Table 9: Comparisons between patient groups at initial assessment and re-assessment.