Adapting the Asthma Life Impact Scale (ALIS) for use in Southern European (Italian) and Eastern European (Russian) cultures

Abstract

Background:

The Asthma Life Impact Scale (ALIS) is a disease-specific measure used to assess the quality-of-life of people with asthma. It was developed in the UK and US and has proven to be acceptable to patients, to have good psychometric properties, and to be unidimensional.

Objective:

This paper reports on the adaptation and validation of the ALIS for use in representative Southern European (Italian) and Eastern European (Russian) languages.

Methods:

The ALIS was translated for both cultures using the dual-panel process. The newly translated versions were then tested with asthma patients to ensure face and content validity. Psychometric properties of the new language versions were assessed via a test–re-test postal survey conducted in both countries.

Limitations:

It is possible that some cultural or language differences still exist between the different language versions. Further research should be undertaken to determine responsiveness. Further studies designed to determine the clinical validity of the Italian ALIS would be valuable.

Results:

Linguistic nuances were easily resolved during the translation process for both language adaptations. Cognitive debriefing interviews (Russia n = 9, male = 11.1%, age mean (SD) = 55.4 (13.2); Italy n = 15, male = 66.7%, age mean (SD) = 63.5 (11.2)) indicated that the ALIS was easy to read and acceptable to patients. Psychometric testing was conducted on the data (Russia n = 61, age mean (SD) = 40.7 (15.4); Italy n = 71, male = 42.6%, age mean (SD) = 49.5 (14.1)). The results showed that the new versions of the ALIS were consistent (Russian and Italian Cronbach’s alpha = 0.92) and reproducible (Russian test–re-test = 0.86; Italian test–re-test = 0.94). The Italian adaptation showed the expected correlations with the NHP and the Russian adaptation showed strong correlations with the CASIS and CAFS and weak-to-moderate correlations with %FEV1 and %PEF. In both adaptations the ALIS was able to distinguish between participants based on self-reported general health, self-reported severity, and whether or not they were hospitalized in the previous week.

Keywords::

• Adaptation
• ALIS
• Asthma
• Patient-reported outcomes
• Quality-of-life

Introduction

Asthma is a chronic respiratory disease characterized by variable and recurring symptoms and reversible airway constriction. External triggering agents cause flares in which the airways narrow and become inflamed. This produces symptoms that include wheeze, breathlessness, chest tightness, cough, and airflow obstruction1. It is estimated that ∼300 million people worldwide have asthma2. As asthma is related to urbanization it is estimated that by 2025 a further 100 million people will become affected2. Asthma accounts for one of every 250 deaths worldwide3. Unsurprisingly, asthma has a considerable economic impact. In the UK it has been estimated that the cost due to lost work days, social security payments, and National Health Service costs is over £2.3 billion a year4.

Measures of symptoms and functioning fail to assess the full impact of asthma adequately. Asthma has a negative impact on a number of other aspects of life such as social relationships5,6, sex7, employment8, and academic study9. Asthma is also associated with a number of psychological problems such as anxiety, depression, and panic disorder10,11. A recent review indicated that even patients with mild illness can have poor quality-of-life (QoL)12. In order to account for such impacts it is important that outcome measures are available that assess the full QoL impact of asthma.

A recent review acknowledged limitations in several of the available patient reported outcome measures (PROs) available for use in asthma studies13. For example, none of the PROs identified were based on a clear theoretical model. In addition, the content of the PROs were mainly concerned with symptoms and functioning, with too few items specific to QoL issues. The Asthma Life Impact Scale (ALIS) is a disease-specific QoL measure that was developed recently to address many of the weaknesses of the other available measures14. Unlike many of the previous measures, it is based on a clear theoretical model—the needs-based model of QoL15,16. This model asserts that QoL is dependent on an individual’s ability to fulfil fundamental needs and that QoL is high when these needs are met. The content of the ALIS was generated directly from asthma patients and different samples were involved in each of the subsequent stages of the instrument’s development. The ALIS is the only measure for asthma that has been developed using Item Response Theory (IRT), widely acknowledged as a more powerful psychometric approach to scale development than classical test theory approaches17–19.

The ALIS was developed simultaneously in the UK and the US14. The purpose of the present study was to report on the adaptation of the ALIS for use in Russia and Italy. It is important that any language adaptation of a measure is based on a thorough adaptation procedure. Language contains many nuances and phrases that, although well understood in the language in which the instrument was developed, are not always clear to non-native speakers. Consequently, it is inappropriate to produce a new language version of a questionnaire by simply translating the content (literal translation).

Adaptations of the ALIS follow a three-stage adaptation procedure. Stage 1 involves translating the measure using the dual panel methodology20. This approach involves conducting two translation panels; a professional panel (to provide the initial translation in the target language) and a lay panel (where items are assessed for comprehension and ‘naturalness’ of language). The second stage involves conducting cognitive debriefing interviews to test the face and content validity of the new measures with patients. The final stage involves conducting a validation survey to assess whether the new measures show appropriate levels of reliability and construct validity.

Methodology

Three main adaptation stages were conducted in each country:

• Stage 1: Translation using the dual panel methodology.

• Stage 2: Cognitive debriefing for face and content validity.

• Stage 3: Assessment of the psychometric properties of the translated measure.

Participants

Participants in the cognitive debriefing and validation stages of the adaptations in each country had a clinical diagnosis of asthma. Due to the nature of the study (validation of questionnaires) full ethics approval was not required. No personally identifiable data was recorded. Patients were provided with information about the study and were informed of their right to withdraw at any time. Patients were included in the study if they were over the age of 18, had a diagnosis of asthma, and did not have significant comorbidities. Patients were recruited between February and September 2010 in the Italian study and between February 2007 and June 2008 in the Russian study. Patients were recruited from the following centres on an opportunistic basis:

• Italy: were recruited from a number of sources including general practitioners and hospitals in Trento, Verona, Padova, and Milan.

• Russia: both inpatients and outpatients were recruited from the Department of Pulmonology, Ida-Viru Keskhaigla (East-Viru Central Hospital), Kohtla-Järve, Estonia. Participants were excluded from the Russian adaptation if they spoke Estonian.

Stage 1: Translation

Participants in the translation panels did not have asthma as the purpose was to ensure the appropriate wording of items in the new language rather than to comment on the content of the questionnaire. The duel-panel methodology was employed to translate the questionnaires in each country20. The dual-panel approach requires that the verification and evaluation of acceptability of translations rests with people who are typical of the patients who will later be asked to complete the questionnaire.

The translation method consisted of conducting professional and lay panels in each country. The panels consisted of between four and seven participants and both panels in each country were chaired by the same group leader. His/her role was to encourage the panel members to reach consensus on the appropriate translations for the instructions, items, and response options. The focus throughout was for participants to produce translations that were conceptually equivalent to the original. The leader was also required to ensure that no panel member was too dominant. Participants worked as a team and all were given the opportunity to express their opinion. One of the ALIS developers also attended the professional panels to explain the precise conceptual meaning of the items to panel members. It is also the job of the instrument developer to help resolve any instances where there are difficulties translating a particular phrase or item.

The professional panels worked predominantly in the target language. Items were presented to the groups one-by-one and their meaning explained. Alternative translations suggested by individual group members were considered by the whole group. Each item was discussed until agreement was reached. Where consensus could not be reached alternative versions of the item were sent for consideration by the lay panel.

The lay panels worked only in the target language. Individuals were selected for this panel if they had an average or lower than average educational level. The purpose of this second panel was to ensure that the final wording of items was appropriate for the average future respondent. Participants were presented with the translation(s) made by the professional panel and asked to comment on it/them in terms of comprehension and acceptability. In particular, they were asked to decide whether phrasing and language were acceptable or whether changes in wording should be made to make the items more ‘natural’ while maintaining their original meaning.

Stage 2: Cognitive debriefing interviews

Interviews were conducted with 15 Italian patients and nine Russian speaking patients to test face and content validity. Specifically, the applicability, comprehensiveness, and relevance of the translated items were investigated. Interviewees were asked to complete the questionnaire in the presence of an interviewer who noted any obvious difficulties or hesitation over individual items. Interviewees were then asked to comment on the questionnaire items, instructions, and response format. Respondents were also asked whether any aspects of their experience of asthma had been omitted.

Stage 3: Validation survey

Validation data were collected by means of a postal survey in Italy and at clinical centers for the Russian adaptation. Participants completed the questionnaire packs at the same location on each assessment. The assessments of validity made in each country varied due to the availability of comparator instruments and the specific interests of the local investigators. The local investigators were required to:

• Conduct a test–re-test survey involving two administrations of the ALIS with a period of 2 weeks between administrations.

• Employ an appropriate comparator measure on the first administration. The Nottingham Health Profile (NHP)21 was used in Italy. For the Russian language adaptation the COPD and Asthma Sleep Impact Scale (CASIS)22 and the COPD and Asthma Fatigue Scale (CAFS)23 were used as comparator measures.

• Collect demographic information on the first administration.

• Collect information on disease severity and overall health status at both assessments. In both countries, self-perceived general health (very good/good/fair/poor) and asthma severity (very severe/quite severe/moderate/mild) were assessed. For the Russian adaptation physician assessment of general health (excellent/very good/good/fair/poor) and severity (very severe/quite severe/moderate/mild) were also assessed.

• Collect relevant clinical information if available. Spirometry assessments of forced expiratory volume (FEV1) and peak expiratory flow (PEF) were available for the Russian adaptation. FEV1 is the volume of air that can be blown out of the lungs in one second and this is expressed as a percentage based on normative data for the person’s age and gender (%FEV1) in the present study. PEF is the mean flow (or speed) of air coming out of the lungs during a period of forced expiration. It is also expressed in the present study as a percentage based on normative data for the person’s age and gender (%PEF).

Outcome measures

The ALIS contains 22 items and has response options of ‘true’ (scored 1) and ‘not true’ (scored 0). Respondents are asked to indicate whether or not their needs are currently being influenced by asthma or its treatment. Higher scores represent greater impairment of QoL.

The NHP consists of six sections; energy level, pain, emotional reactions, sleep, social isolation, and physical mobility. Each item, which enquires whether or not the problem currently applies to them, is scored I for a ‘yes’ response and 0 for a ‘no’ response. Section scores are linearly transformed so that scores range from 0 (no problems) to 100 (severe problems). A unidimensional index of impairment—the NHP-D—can also be generated from a sub-set of NHP items24. High scores on the NHP sections indicate worse health status.

The CASIS is a 7-item PRO that measures sleep disruption associated with either asthma or COPD. Item responses are summed to form a total raw score. These are linearly transformed into a 0–100 total score scale with higher scores indicating a greater impact of asthma on sleep. The CAFS is a 12-item PRO that assesses fatigue related to asthma or COPD. Item responses are scored in the same way as for the CASIS, with higher scores indicating greater levels of fatigue. The recall period for both the CASIS and CAFS is the previous week.

Statistical analyses

Non-parametric statistical tests were used throughout the analyses due to the ordinal nature of the measures employed. Mean and SD scores are also provided in the results section to assist with the interpretation of the data. As the purpose of the surveys was to test the validity of the ALIS, individuals with missing data on any measure were excluded from the analyses.

Internal consistency was assessed using Cronbach’s alpha coefficients. Alpha measures the extent to which items in a scale are inter-related. A low alpha (below 0.7) indicates insufficient relations between the items25. In addition each item is correlated with the total score (corrected-item total coefficients (CITCs)). If this correlation is low (below 0.2) it can indicate that the item is not contributing adequately to the overall scale. If the correlation is high (above 0.8) it suggests that the item is redundant, adding little extra information to the scale.

The test–re-test reliability of a measure is an estimate of its reproducibility over time when no change in condition has taken place. A high correlation indicates that the instrument produces low levels of random measurement error. A minimum Spearman rank correlation coefficient of 0.85 is required26.

ALIS scores were compared for males and females and for respondents who were above or below the median age. Significant differences in ALIS scores between groups split by gender and age group could suggest gender or age bias in the items. Mann-Whitney U-tests were employed to test for differences between the groups.

Convergent validity was evaluated by assessing the level of association between scores on the adapted scale and those on related sections of the comparator scale. In the Italian adaptation, the NHP-D was expected to correlate more strongly with the ALIS as overall distress is a similar construct to QoL. Stronger correlations were also expected with the NHP-Emotional reactions, Energy, Sleep, and Physical mobility, as these are all related to QoL. In the Russian adaptation moderate-to-strong correlations were expected with the CASIS and CAFS, as these measure asthma-specific sleep and fatigue issues. Spearman rank correlation coefficients were employed to test the level of the associations.

Known group validity was assessed by testing the ability of the ALIS to distinguish between groups of people that differ according to a factor that is known, or suspected, to influence scores. ALIS scores were predicted to be lower for respondents who;

• rated their general health better,

• rated their asthma severity as less severe,

• had lower levels of hospitalizations in the previous week,

• were not receiving treatment for their asthma,

• were rated by clinicians to have less severe asthma, and

• had higher %PEF and %FEV1 values (above and below the median).

Results

Stage 1: Translation

Professional panel members (Russia n = 6; Italy n = 7) had high levels of educational achievement with most holding a University degree. Lay panel members (Russia and Italy n = 5) were monolingual speakers in the target language and had lower levels of educational achievement.

The Russian professional panel was able to translate the ALIS with little difficulty. Literal translations were suitable for several items. Some English idioms created more detailed discussion when there was no literal translation for the phrase. For example, item 5 ‘I have to pace myself’ was translated as ‘Мне нужно рассчитывать свои силы’ (I have to calculate my strength (capability)). In these cases the instrument developer checked the translations carefully with the group to ensure that correct meaning was maintained. Overall the panel progressed without major difficulties and it was not necessary to provide the lay panel with alternative translations for any of the items. Lay panel members agreed with the translation of the instructions and with most of the item translations. Only two translations were queried. Panel members had difficulty understanding the translation of the word ‘affect’ in the items My self-confidence is affected and My asthma affects my close relationships. The translation of ‘affect’ has two different meanings in Russian and could have been easily misinterpreted. Consequently, panel members suggested revised translations for ‘affect’ that emphasized impairment. Only minor changes were needed for the other items before the questionnaire was ready for assessment by the cognitive debriefing interviewees.

The Italian professional progressed well but detailed discussions were needed for some items. The professional panel found that literal translations were only appropriate for a couple of items. Difficulties with the translations mainly arose from the lack of equivalent Italian expressions for the colloquial English phrases. Panel members discussed at length several items that proved problematic to translate into Italian. For example, no equivalent Italian phrase was available for ‘let down’ in the item I feel like I let other people down. Three different alternatives were provided for this item, each of which conveyed the impression of disappointing others’ expectations. Where difficulties were found in translating the items the ALIS developer was able to assist the group in arriving at translations that captured the same concept as the original. Several alternatives were suggested for some of the items and these were sent to the lay panel for consideration. This panel was able to select the options that were clearer and easier to understand. Minor alterations to existing translations were suggested for several items.

Stage 2: Cognitive debriefing

Interviewee details are shown in Table 1. Patients in the Russian adaptation completed the ALIS in a mean of 3 (range 2–4) minutes. Patients in the Italian adaptation completed the questionnaire in a mean of 4.2 (range 3–6) minutes. In both language adaptations, all patients found the questionnaire clearly expressed and easy to understand. Participants felt that the items were relevant and they were able to relate to the ideas expressed. In the Italian study all of the participants felt that the items covered all the important areas connected to living with asthma and that no aspect of their experience had been omitted. In the Russian study one patient thought that the questionnaire should contain information on work environment hazards and another patient thought information on inability to keep pets should be added. These issues were considered to be idiosyncratic. All the other Russian patients thought that no aspect of their experience had been omitted. Only minor changes were made to either the Russian or Italian adaptations of the ALIS following the cognitive debriefing interviews.

Table 1.  Demographic information for the cognitive debriefing and validation samples.

	Cognitive debriefing sample		Validation sample
	Russia (n = 9)	Italy (n = 15)	Italy (n = 61)	Russia (n = 71)
Gender (%)
Male	1 (11.1)	10 (66.7)	26 (42.6)	31 (43.7)
Female	8 (88.9)	5 (33.3)	34 (55.7)	40 (56.3)
Missing	0	0	1 (1.6)	0
Age (years)
Mean (SD)	55.4 (13.2)	63.5 (11.2)	40.7 (15.4)	49.5 (14.6)
Median (IQR)	56.0 (45.5–69.0)	67.0 (53.0–70.0)	38.0 (26.5–50.0)	53.0 (41.5–58.5)
Range	31.0–70.0	45.0–83.0	17.0–76.0	15.0–71.0
Marital status (%)
Married/living as married	6 (66.7)	7 (46.7)	29 (47.5)	43 (60.6)
Widowed	3 (33.3)	5 (33.3)	3 (4.9)	10 (14.1)
Divorced/separated	0	0	7 (11.5)	6 (8.5)
Never married	0	3 (20.0)	22 (36.1)	12 (16.9)
Employment status (%)
Full-time	4 (44.4)	4 (26.8)	36 (59.0)	32 (45.1)
Part-time	0	0	7 (11.5)	3 (4.2)
Homemaker	0	4 (26.6)	5 (8.2)	1 (1.4)
Retired	3 (33.3)	7 (46.7)	6 (9.8)	19 (26.8)
Long term sick leave	2 (22.2)	0	0	11 (15.5)
Student	0	0	7 (11.5)	2 (2.8)
Other	0	0	0	3 (4.2)

IQR, Inter-quartile range.

Stage 3: Validation

Descriptive statistics

Demographic information for the validation samples are shown in Table 1. In both the Russian and the Italian samples ∼56% of the samples were female. The mean age of the Italian sample was 38 (range 17–76) years compared with a higher mean age of 53 (range 15–71) years for the Russian sample. Both samples consisted mostly of participants who were married or living as married.

Table 2 shows disease information for the two samples. Both the Russian and the Italian samples had a median illness duration of 15 years. The clinical assessments indicated that the Italian sample was milder than the Russian one. Half of the Italian sample reported their asthma to be mild and 44% reported their general health to be good. In contrast, 28% of the Russian sample reported their general health to be good and only 12.7% reported their asthma severity to be mild.

Table 2.  Patient-reported disease information.

	Italy (n = 61)	Russia (n = 71)
Asthma duration (years)
Median (IQR)	15.0 (10.0–20.0)	15.0 (10.0–20.5)
Mean (SD)	16.3 (10.0)	16.7 (8.6)
Range	44 (1–45)	39 (1–40)
Self-reported general health (%) Time 1
Very good	6 (9.8)	4 (5.6)
Good	27 (44.3)	20 (28.2)
Fair	23 (37.7)	34 (47.9)
Poor	5 (8.2)	13 (18.3)
Very poor	0	0
Self-reported asthma severity (%) Time 1
Mild	31 (50.8)	9 (12.7)
Moderate	24 (39.3)	32 (45.1)
Quite severe	6 (9.8)	27 (38.0)
Very severe	0	2 (2.8)
Missing	0	1 (1.4)
Hospitalized during last week (%)
Yes	11 (18.0)	51 (71.8)
No	50 (82.0)	19 (26.8)
Missing	0	1 (1.4)
Receiving treatment for asthma (%)
Yes	24 (39.3)	68 (95.8)
No	34 (55.7)	1 (1.4)
Missing	3 (4.9)	2 (2.8)

IQR, Inter-quartile range.

Additional clinical data for the Russian sample are shown in Table 3.

Table 3.  Russian validation: Clinical validation data at Time 1.

	Russia (n = 71)
%FEV1
Median (IQR)	71.2 (60.4–82.9)
Mean (SD)	73.0 (18.8)
Range	41–138
%PEF
Median (IQR)	61.5 (49.8–74.1)
Mean (SD)	63.2 (18.4)
Range	21–116
Clinician-rated general health (%)
Very good	11 (15.5)
Good	16 (22.5)
Fair	28 (39.4)
Poor	14 (19.7)
Very poor	2 (2.8)
Clinician-rated asthma severity (%)
No symptoms	4 (5.6)
Mild	22 (31.0)
Moderate	27 (38.0)
Quite severe	16 (22.5)
Very severe	2 (2.8)

IQR, Inter-quartile range.

Questionnaire descriptive scores

Minimal levels of missing data were observed. In the Italian adaptation no missing ALIS data was found at Time 1. At Time 2 two patients dropped out (3.28% of sample) but no missing ALIS data was observed for the other patients. In the Russian adaptation one patient at Time 1 missed items. At Time 2, two patients dropped out (2.8% of sample) and four patients missed items.

ALIS scores at the Time 1 and Time 2 assessments are shown in Table 4. Russian ALIS scores at Time 1 were significantly higher than Italian scores (p < 0.01) supporting the other clinical information. Minimal levels of floor and ceiling effects were found in the Russian data. The Italian sample had high floor effects at both time points supporting the mild nature of the sample.

Table 4.  Descriptive statistics for the ALIS.

	Median (IQR)	Mean (SD)	Range	% scoring minimum	% scoring maximum
Time 1
Italy	5 (2–9)	5.8 (5.4)	0–22	21.3	1.6
Russia	8 (3.8–13)	8.3 (5.7)	0–22	4.2	1.4
Time 2
Italy	4 (1–8)	5.3 (5.1)	0–21	21.3	0.0
Russia	7 (2.5–12)	7.7 (5.9)	0–22	7.0	1.4

IQR, Inter-quartile range.

Internal consistency

Cronbach’s alpha coefficients were 0.92 in both adaptations, indicating adequate inter-relatedness of items. Corrected item total correlations (CITC) and Cronbach’s alpha if each item is deleted is shown in Table 5. All CITC coefficients were between 0.2–0.8 in the Italian adaptation. Only one item in the Russian sample had an ITC out of the desired 0.2–0.8 range (item 18 ‘I worry that other people will think I’m lazy’). The single low ITC score is unlikely to compromise the overall reliability of the scale. In addition, the deletion of items in either adaptation did not substantially alter the overall Cronbach’s alpha coefficient.

Table 5.  Corrected item total correlations.

Item	Italian		Russian
	Corrected item total correlation	Cronbach’s alpha if item deleted	Corrected item total correlation	Cronbach’s alpha if item deleted
Inability to be adventurous	0.62	0.91	0.59	0.91
Dependent on treatment	0.40	0.91	0.37	0.92
Social limitations	0.50	0.91	0.49	0.92
Feel old	0.52	0.91	0.45	0.92
Need to pace activities	0.67	0.91	0.50	0.92
Self-confidence	0.46	0.91	0.45	0.92
Preoccupation with medication	0.55	0.91	0.55	0.91
Limit activities	0.75	0.91	0.71	0.91
Let others down	0.39	0.91	0.57	0.91
Cautious where to go	0.55	0.91	0.64	0.91
Affects relationships	0.49	0.91	0.42	0.92
Lowered mood	0.65	0.91	0.71	0.91
Interest in sex	0.63	0.91	0.41	0.92
Prevents visiting	0.47	0.91	0.69	0.91
Limits potential	0.59	0.91	0.59	0.91
Feel isolated	0.64	0.91	0.55	0.91
Need for energy	0.65	0.91	0.67	0.91
Appear lazy	0.35	0.92	0.13	0.92
Planning activities	0.58	0.91	0.59	0.91
Preoccupation with asthma	0.57	0.91	0.75	0.91
Controlled by illness	0.58	0.91	0.76	0.91
Limit social activities	0.70	0.91	0.58	0.91

Test–re-test reliability

The test–re-test correlation coefficients were 0.86 (n = 59) for Italy and 0.94 (n = 65) for Russia. Both values indicate that the adaptations have good reproducibility and would produce low levels of measurement error. This is supported further by the stability of scores between assessments shown in Table 4.

Association with demographic factors

There were no significant differences in mean ALIS scores, between participants grouped by gender or age (above vs below the median age) in either the Russian or the Italian samples.

Convergent validity

Table 6 shows the correlations between the ALIS and the other outcome measures for both adaptations. In the Italian adaptation the ALIS had moderate correlations with Energy level, the NHPD, and Physical mobility. The ALIS had the weakest correlation with the social isolation and pain sections of the NHP.

Table 6.  Convergent validity between ALIS and other outcome assessments.

	Italy		Russia
	Correlation	n	Correlation	n
NHP Energy level	0.63	60
Pain	0.40	60
Emotional reactions	0.46	60
Sleep	0.46	60
Social Isolation	0.34	59
Physical mobility	0.52	59
NHPD	0.56	59
CASIS			0.85	70
CAFS			0.87	68
%FEV1			−0.35	70
%PEF			−0.40	70

All correlations significant, p < 0.01.

In the Russian adaptation the ALIS had strong correlations with both the CAFS and CASIS. The ALIS had weak-to-moderate correlations with %FEV1 and %PEF.

Known group validity

There were no significant differences in ALIS scores by duration of asthma (above vs below the median value) in the Italian adaptation. However, this difference was statistically significant in the Russian adaptation.

Known group validity results for both adaptations are shown in Table 7. Both adaptations were able to distinguish between participants based on their self-perceived general health and severity of asthma. Individuals with worse health and greater severity had significantly higher ALIS scores. Both adaptations were able to distinguish between participants based on whether or not they required hospitalization in the previous week. Those individuals that required hospitalization scored higher on the ALIS.

Table 7.  Mean ALIS scores by patient reported known groups (Time 1).

Variable	Patient sub-group	Italy		Russia
		n	Mean (SD)	n	Mean (SD)
Asthma duration	<15 years	28	4.6 (4.3)	30	6.5 (5.6)
	≥15 years	33	6.9 (6.0)	39	9.5 (5.6)
	p	61	0.13	69	<0.05
Self-perceived general health	Very good/good	33	2.6 (2.8)	23	3.3 (3.0)
	Fair	23	8.1 (3.9)	34	8.9 (4.3)
	Poor/very poor	5	16.6 (4.7)	13	15.5 (3.9)
	p	61	<0.001	70	<0.001
Self-perceived COPD severity	Mild	31	3.3 (3.2)	9	3.6 (4.2)
	Moderate	24	6.6 (4.2)	32	6.2 (4.4)
	Quite severe	6	15.8 (6.3)	29	12.1 (5.2)
	p	61	<0.001	70	<0.001
Hospitalized during previous week?	Yes	11	11.6 (6.7)	51	9.4 (5.5)
	No	50	4.6 (4.1)	19	5.3 (5.3)
	p	61	<0.001	70	<0.01
Receiving asthma treatment?	Yes	24	8.0 (6.2)
	No	34	4.6 (4.5)
	p	58	<0.05

The Italian adaptation was able to distinguish between participants based on whether or not they received treatment for their asthma (p < 0.05). All but one of the Russian respondents were receiving treatment for their asthma.

Known group validity results for the clinical data available for the Russian validation study are shown in Table 8. The Russian ALIS was able to distinguish between %FEV1 and %PEF (above vs below the median value). Those patients with lower %FEV1 and %PEF had higher ALIS scores indicating poorer QoL. The ALIS was also able to distinguish between participants based on physician-assessed general health and severity. Those categorized by the physician as having poorer health and those with greater severity of asthma scored higher on the ALIS.

Table 8.  Mean ALIS scores by clinician-rated known groups (Russia time 1).

Variable	Patient sub-group	Russia
		n	Mean (SD)
Clinician-rated general health	Excellent/very good	11	2.6 (2.8)
	Good	16	5.5 (4.3)
	Fair	27	8.5 (4.5)
	Poor/very poor	16	14.6 (4.1)
	p	70	<0.001
Clinician-rated asthma severit	No symptoms/mild	26	3.7 (3.2)
	Moderate	26	9.2 (4.4)
	Quite severe/very severe	18	13.7 (4.8)
	p	70	<0.001
%FEV1	<Median	34	10.1 (5.9)
	>Median	36	6.6 (5.1)
	p	70	<0.01
%PEF	<Median	34	11.2 (5.2)
	>Median	36	5.5 (4.7)
	p	70	<0.001

Discussion

The ALIS is a new QoL measure, based on the needs-based model of QoL, which examines the everyday impact of asthma. Its content was generated by means of interviews with asthma patients and measures how they are affected from their own perspective. The ALIS was developed in parallel in the UK and US. It has subsequently been translated into Belgian-Dutch and Belgian-French, Canadian-English and Canadian-French, Danish, Dutch, French, German, Swedish, Turkish, Spanish, and US-Spanish. This article describes the adaptation of the ALIS for use in Italy and Russia using a three-stage adaptation procedure. This approach has been successfully applied in the adaptation of several instruments that have applied the needs-based model of quality-of-life27–29. The strength of the three-stage adaptation procedure over other approaches is that the adaptations go through several tiers of critical scrutiny.

The results of the translation stage showed that the ALIS could be successfully translated for use in representative Eastern and Southern European languages with few problems. Greater discussion was required for items that included English idioms or colloquial phrases. However, suitable conceptual translations were found for all items in both languages. The addition of a lay panel ensures that the level of language used in the final measure is appropriate for the average respondent in the new culture.

The cognitive debriefing interviews indicated that the content of the questionnaires was appropriate in terms of content, comprehension, and ease of completion. Few changes in language were necessary as a result of this stage of the adaptation as patients found the content of the questionnaire to be suitable, easy to understand, and comprehensive.

The validation surveys showed that both measures had high levels of internal consistency and reproducibility, which, in turn, suggested that the adaptations would be responsive. The adapted questionnaires were able to distinguish between participants based on a number of different factors providing evidence of the scales’ construct validity. Overall, both adaptations showed expected relations with the comparator instruments which assess similar constructs. A difference was observed between the two adaptations in the degree of relationship between the ALIS and the sleep and fatigue scales used in each study. In the Russian study strong correlations were observed between the ALIS and sleep and fatigue scales and in the Italian study the relationships were moderate. These differences may be due to the different measures employed in the two studies. The NHP energy and sleep scales are generic measures and contain only three and five items, respectively. In contrast the CAFS and CASIS contain more items and were developed specifically to assess sleep and fatigue issues in asthma and COPD patients.

QoL is not just a description of patients’ health status but is a reflection of the way in which they perceive and react to their health status and to other non-medical aspects of their lives30. Consequently it was interesting to find that, in the Russian adaptation, moderate correlations were observed between the ALIS and the %FEV1 and %PEF assessments. Generally, low cross-sectional associations are found between asthma-specific patient reported outcome measures and spirometry assessments13,31. The Russian ALIS sample was more severe than that in Italy, probably because they were recruited exclusively via hospitals while the Italian sample was recruited from both general practice and hospitals. It is possible that the major symptom of asthma, breathlessness, has a strong association with the QoL of patients who are severely affected by their asthma.

Study limitations

Several limitations were identified with the study. First, the Russian adaptation was carried out with ethnic Russians living in Kohtla-Järve, Estonia. However, the movement of Russians into Estonia was quite a recent event historically (post-World War II) and this population lives separately from the native people and continues to speak Russian rather than Estonian. Patients were also excluded from the study if they spoke Estonian. Russia is a large country and the language spoken varies from region to region. It would be beneficial to run a lay panel in any part of Russia prior to using the ALIS.

The present study has not determined the responsiveness of the ALIS. However, the high test–re-test scores (indicating low levels of measurement error) and strong evidence of known group validity suggest that the new scales may be responsive to change. Further research should be undertaken to determine responsiveness.

Clinician-rated severity and pulmonary function data were not available for the Italian adaptation. Further studies designed to determine the clinical validity of the Italian ALIS would be valuable.

The original development of the ALIS involved the use of Rasch analysis to identify a unidimensional scale. It would have been desirable to re-assess the measurement properties of the ALIS using Rasch analysis in the new languages. However, the relatively small sample sizes meant that such analyses were likely to lead to findings that would lack robustness.

Finally, although the adaptations were successful and the results appeared comparable to the original UK version, it is possible that some cultural or language differences still exist between the different language versions. Although rarely investigated by researchers it is possible that such cultural differences could affect the results of international clinical trials and studies where data from different countries are combined. The application of Differential Item Functioning (DIF) using Rasch analysis is particularly effective in investigating whether cultural and language issues affect participants’ responses to items32–34.

Conclusions

The ALIS was successfully adapted into Russian and Italian. Few problems were experienced in producing conceptually equivalent items. The new translations demonstrated good convergent/divergent validity, reproducibility, and construct validity and should prove valuable in determining the benefits of treatment. The present study has shown that the measure can be successfully adapted for use in Eastern- and Southern-European cultures with the adaptations demonstrating similar psychometric properties to the original UK and US versions.

Transparency

Declaration of funding

This study was funded by Novartis AG. The article was prepared without restrictions from the sponsors. All authors meet criteria for authorship as recommended by the International Committee of Medical Journal Editors (ICMJE) and all authors contributed to the development of the manuscript.

Declaration of financial relationships

JT, SPM, and SRC were all employees of Galen Research Ltd. at the time that the manuscript was prepared. MT was an employee of the University of Tartu and NCO was an employee of the University of Verona when the manuscript was prepared.

Acknowledgments

The authors would like to acknowledge the help of the Italian and Russian patients who participated in the study. They would also like to thank Alastair Glendenning for his advice and encouragement during the study.