reliability and validity of the german version of the ...

I

Psychology and Health . 1997, Vo1.12, pp. 149-159 Reprints available directly from the Publisher Photocopying permitted by license only

© 1997 OPA (Overseas Publishers Association) Amsterdam B.Y. Published in The Netherlands by Harwood Academic Publishers Printed in Malaysia

RELIABILITY AND VALIDITY OF THE GERMAN

VERSION OF THE SICKNESS IMPACT PROFILE

IN PATIENTS WITH CHRONIC OBSTRUCTIVE

PULMONARY DISEASE

B. O. HÜITER I and G. WÜRTEMBERGER2 lDepartment of Neurosurgery, Technical University (RWTH) Aachen, FRG 2IIl. Medizinische Klinik, Carl-Thiem Klinikum, Cottbus, FRG (Reeeived 17 August, 1994; in final form 4 Getober, 1995)

Reliability and validity of the German version of the SickneS~ Impact Profile (SIP) were investigated in a sampie of 58 O,-dependent patients with chronic obstructive pulmonary disease (COPD). Internal consistency (Cronbach Alpha) was .93 for the entire instrument, while it was .85 for the SIP psycho-social scale and .89 for the SIP physical scale. The Cronbach Alphas ranged from .55 to .86 for the subscales of the German SIP ·version. The subscale "Eating" was discarded because of low internal consistency (.40) and split-half reliability (.24). Indices of respiratory function (FEV" FVC, PaO" SaO,), functional impairment and dyspnea (BDI) were substantiaUy associated with the total SIP, psychosocial and physical score, ranging from r = -.33 to r = -.48. A stepwise discriminant analysis between the 14 patients, who died during the course ofthe study and the survivors revealed one single discriminant function with an Eigenvalue of .5859 (canonical correlation r = .61 ; Chi?' = 21.21; df = 3; p < 0.001 ). The FVC (Wilk 's Lambda .76), the severity of dyspnea (BDI) (Wilk 's Lambda .71 ), the SIP total score (Wilk' s Lambda of .65) and the FEV, (Wilk's Lambda .63) were identified as discriminating variables, allowing the correct classification of 85% of the survivors and 91 % of the fatalities . Keywords: Quality of life, Sickness Impact ProfIle, German version of the Sickness Impact Profile, Psychometric analysis, Chronic obstructive pulmonary di sease, Liquid oxygen therapy.

INTRODUCTION In modem medicine, quality of life has become an important criterion for assessing disease impact and treatment outcome (Aaronson, 1988; Van Dam and Aaronson , 1988; Hütter, 1990; Najman and Levine, 1981; Ware, 1984). For assessing quality of life, the use of current rating scales like the Glasgow Outcome Scale (Hütter and Gilsbach, 1993; Jennett and Bond, 1975; Jennett, Snoek, Bond and Brooks, 1981) or the Kamofsky Performance Status Scale (Hutchinson, Boyd, Feinstein, Gonda, Hollomby and Rowat, 1979; Yates, Chalmer and McKegney, 1980) is insufficient, because they on l.y assess the physical-functional level of the patients. Therefore, they do not cover all relevant aspects of quality of life. In modem behavioural medicine, a multidimensional concept of quality of life is used (Aaronson, 1988; F1etcher, Hunt and Bulpitt, 1987; Hütter, 1990; Van Dam and Aaronson, 1988; Ware, 1984; Würtemberger, Hütter, Hirsch and Matthys, 1992). The Sickness Impact Profile (SIP), as a self-report measure of quality of life, was developed according to a psychosocial concept of quality of life which is assessed on a concrete behaviourallevel. It has proven reliability and validity Address for correspondence: B. 0 . Hütter, Ph. D. Departrnent ofNeurosurgery, Technical University, (RWTH) Aachen, 52057 Aachen, Germany, Pauwelsstr. 30. Phone 0241 /8088480; Fax 0241 /8888420.

149

•

150

B. 0. HÜTTER AND G. WÜRTEMBERGER

(Bergner, Bobbitt, Pollard, Martin and Gilson, 1976a,b; Bergner, Bobbitt, Carter and Gilson, 1981; Follick, Smith and Ahern, 1985; Jones, Baveystock and Littlejohns, 1989; McSweeny, Grant, Heaton, Prigatano and Adams , 1985). In the present study, the internal consistency and split-half reliability of the German version of the SIP (Hütter, 1990) were investigated in patients with COPD and hypoxemia. It was the aim to corroborate the convergent validity of the German version of the SIP by means of studying the associations with different clinical and physiological indices of lung function and relevant subscales of the German personality form Freiburger Personality Inventory (FPI-R).

PATIENTS AND METHODS From January 1991 to June 1991, a total of 132 patients with COPD and hypoxemia were consecutiyely treated with liquid oxygen therapy (LOX) in the Department of Pneumonology of the University of Freiburg. The patients suffered from severe end stage COPD fulfilling the NOIT criteria (Benson,. 1987; Nocturnal Oxygen Therapy Trial group, 1980). From May to June 1991 , all patients were invited by mail to a medical follow-up examination. A set of questionnaires including the German version of the Sickness Impact Profile (SIP) was added to the letter. The subjects were ' asked to fill out the questionnaires at home and to send them back to the Department of Pneumonology. Outbf the 132 patients, 8 (6%) had died in the meantime and 15 (11 %) had moved to an unknown address. A further 11 (8 %) patients had been institutionalized because of hypoxic dementia and/or delirium. Of the remaining 98 patients who could be contacted, 63 (64%) were willing to fill out the questionnaire and to take part in the pneumonological follow-up examination. Five of them had to be excluded because they omitted severa1 items of the SIP. Therefore, the present study includes aseries of 58 hypoxemic patients with COPD. At the time of the study (May/June 1991) all patients had already received LOX for one to five months. Table 1 gives the pneumono1ogica1 and functiona1 data of the patients in the study group.

Table 1

Physiological and clinical characteristics of the patients in the study sampie (n

Age (yrs)

=58)

Mean

SD

Min

Max

54.4

21.4

15.0

76.0

0.6

2.4

FVC (% pred)

29.3

8.5

15.0

45.0

Pa0 2 (mmHg)

56.5

5.6

48.0

64.0

Sa0 2 (mmHg)

84.6

4.2

76.0

92.0

1.6#

FEV! (L/s)

Impairment (BDI) !

2.0#

0.0

3.0

Effort (BDI)!

2.0#

0.0

3.0

Summary score (BDI)!

4.0#

0.0

6.0

# Median

'BOI = Baseline Dyspnea Index (Mahler

e/

al., 1984)

CHRONIC OBSTRUCTIVE PULMONARY DISEASE

151

A total of 47 (80%) COPD patients were retired, 37 (64%) of whom as a consequence of their illness. Out of the 11 (20%) patients who worked regularly, eight did only light work and seven reported that they could only work for a short time and had to rest frequently. Of the 58 participants, 21 (37%) had been treated before with common oxygen concentrators. The mean age was 54.4 years with a range from 15 to 76 years. The 15 year-01d patient was a girl with severe cystic fibrosis. There were 49 (84%) male and nine (16%) female participants. In September 1992, a survey of fatalities was performed in order to compute the one-year survival rate. From May 1991 to September 1992, 14 (24%) patients died of their illness. This number is somewhat higher than the overall mortality of oxygen-dependent COPD patients with a five-year survival rate of about 30% (Benson, 1987). However, the study participants suffered from severe end-stage COPD. Rating

0/ pneumonological /unctional

capacity

The functional impairment and the magnitud~ of effort subscales of the Base1ine Dyspnea Index (BDI) were emp10yed to assess dyspnea and exercise capacity (Mahler, Harver, Rosiello and Daubenspeck, 1989). Validity and reliability of t~e BDI in patients with chronic respiratory disease have been demonstrated by previous studies (Mahler et al. , 1989; Mahler, Rosiello, Harver, Lentine, McGovern and Daubenspeck, 1987; Mahler, Weinberg, Wells and Feinstein, 1984). The ratings of each BDI functional impairment and BDI magnitude of effort score were added to complete a BDI summary score. The functional assessment was performed during the pneumonological follow-up examination by the second author who had at this time no information about the quality of life data of the patients. Assessment

0/ physiological parameters 0/ lung function

The physiological parameters of lung function were assessed during the pneumonological follow-up examination by physicians who were blinded to the quality of life data of the patients. For organisational reasons, the time-delay between the completion of the questionnaires and the assessment of lung function ranged between some days and four weeks. The severity of airway obstruction was quantified by the forced expiratory vo1ume in one second (FEV and the forced vital capacity (FVC). Both parameters were measured in the sitting position using a portable spirometer (Spirometrics model 2451). The FVC was expressed as a percentage of the predicted value. Predicted normal values were adapted from Morris, Koski and Johnson (1971). The Pa0 2 and the Sa02 were analyzed at the same time while the patients were breathing room air. The Pa0 2, arterial carbon dioxide tension (PaCO) and the pH were determined using an AVL 995 blood gas analyzer, allowing the calculation of the Sa0 2 (Würtemberger, Müller, Matthys and Sokolov, 1994). j

Assessment

)

0/ quality 0/ life

For assessing quality of life, the German version of the Sickness Impact Profile (SIP) was employed (Hütter, 1990). The Arnerican version of the SIP is a self-report questionnaire that measures sickness-related complaints in 12 aspects of quality of life: sleep and rest; mobility; horne management; social interaction; ambulation; ernployrnent;

152

B. O. HüTIER AND G. WÜRTEMBERGER

recreation and pastimes; eating; communication; emotional and affective behaviour; body care and movement; alertness (Bergner et al., 1976a,b; Bergner et al. , 1981). It contains 136 items and takes about 20 minutes to complete. Since the items of the SIP are formulated on a concrete behavioural level, the SIP is an objective and verifiable measure of quality of life. In several studies a good convergent and divergent validity of the SIP has been shown (Bergner et al., 1976b; Bergner et al., 1981; Follick et al. , 1985; Iones et al., 1989; McSweeny et al., 1982). The 24-hour retest-reliability is about .90 for the entire instrument, while the internal consistency is about .80 (Bergner et al., 1976a). The SIP was translated into German, item for item, using a consensus procedure by the authors in close collaboration with a coworker of the first author, who speaks North American English, but who has lived now in Germany for several years. In case of divergent translations the disagreement was cleared. In the German translation of the SIP, the "Emotional and affective behaviour", the "Alertness", and the "Body care and movement" subscales were discarded because they showed insufficient reliability and/or validity in an unpublished pilot study (Hütter, 1990). Therefore, the German version of the SIP contains 111 items in nine subscales, which enables the calculation of 'nine subscale impairment scores, a sum score of total impairment, a psychosocial score inc1uding the "Social interaction", "Recreation and pastimes", and "Communication" subscales, and a physical score computed of the "Sleep and rest", "Ambulation", "Mobility", and "Horne-management" scales. In the instructions for each subscale of the SIP the respondents were requested to ans wer the questions as to how they perceived their symptoms on the day of filling out the form and further to report only such complaints which are related, according to their own opinion, to their present physical illness. Assessment of emotional adaptation and personality traits

For assessing the emotional state of the COPD patients, the FPI-R, a German standard personality form was employed (Fahrenberg, Hampel and Selg, 1984). It consists of 12 different bipolar subscales encompassing 138 items with dichotomaus response possibilities (yes/no). The subscales of the FPI-R consist of 11 to 14 items and same items contribute to more than one subscale. The transformation of the raw scores of every subscale into stanine scores was adjusted for age and sex (Fahrenberg, Hampel and Selg, 1984). The FPI-R incorporates personality traits as well as personality disturbances typical for people with somatic complaints. The dimensions of the FPI Rare 1. life-satisfaction vs. life-dissatisfaction, 2. social responsibility vs. selfishness, 3. ambition vs. loss of motivation, 4. social insecurity vs. assertiveness, 5. sensitivity vs. calm, 6. aggressiveness vs. inaggressiveness, 7. strain vs. stability, 8. bodily complaints vs. health, 9. bodily cancern vs. no bodily cancern, 10. openness vs. social desirability, 11. extraversion vs. introversion and 12. emotionallability vs. emotional stability. Ta control for minimizing or social desirability response bias confounding the life quality assessment by the SIP, the social desirability subscale of the FPI-R was employed. Empirical work has demonstrated that this subscale corresponds to such response sets as social desirability and disclosedness (Amelang and Borkenau, 1981). In several studies adequate validity and reliability of the FPI-R subscales could be demotlstrated (Fahren berg et al., 1984).

I

Psyehology and Health , 1997, Vol. 12, pp. 149-159 Reprints available directly from the Publisher Photocopying permitted by license only

© 1997 OPA (Overseas Publishers Association)

Amsterdam B.V. Published in The Netherlands by Harwood Academic Publishers Printed in Malaysia




PULMONARY DISEASE

B. O. HÜTTERI and G. WÜRTEMBERGER2 IDepartment 01 Neurosurgery, Technical University (RWTH) Aachen, FRG 2I/I. Medizinische Klinik, Carl-Thiem Klinikum, Cottbus , FRG (Reeeived 17 August, 1994; in final form 4 OelOber, 1995) Reliability and validity of the German version of the Sickness Impact Profile (SIP) were investigated in a sampie of 58 0 2-dependent patients with chronic obstructive pulrnonary disease (COPD). Interna! consistency (Cronbach Alpha) was .93 for the entire instrument, while it was .85 for the SIP psycho-social scale and .89 for the SIP physical scale. The Cronbach Alphas ranged from .55 to .86 for the subscales of the German SIP ·version. The subscale "Eating" was discarded because of low internal consistency (.40) and split-half reliability (.24). Indices of respiratory function (FEV" FVC, Pa0 2, Sa02), functional impairrnent and dyspn~a (BOI) were substantially associated with the total SIP, psychosocial and physical score, ranging from r = -.33 to r = -.48. A stepwise discriminant analysis between the 14 patients, who died during the course ofthe study and the survivors revealed one single discriminant function with an Eigenvalue of .5859 (canonical correlation r = .61 ; eM' = 21.21; df= 3; p < 0.001). The FVC (Wilk ' s Lambda .76), the severity of dyspnea (BOI) (Wilk's Lambda .71), the SIP total score (Wilk 's Lambda of .65) and the FEV, (Wilk's Lambda .63) were idenlified as discriminating variables, allowing the correct classification of 85% of the survivors and 91 % of the fatalities . Keywords: Quality of life, Sickness Impact Profile, German version of the Sickness Impact Profile, Psychometric analysis, Chronic obstructive pulmonary disease, Liquid oxygen therapy.

INTRODUCTION In modern medicine, quality of life has become an important criterion for assessing disease impact and treatment outcome (Aaronson, 1988; Van Dam and Aaronson, 1988; Hütter, 1990; Najman and Levine, 1981; Ware, 1984). For assessing quality of life, the use of current rating scales like the Glasgow Outcome Scale (Hütter and Gilsbach, 1993; Jennett and Bond, 1975; Jennett, Snoek, Bond and Brooks, 1981) or the Karnofsky Performance Status Scale (Hutchinson, Boyd, Feinstein, Gonda, Hollomby and Rowat, 1979; Yates, Chalrner and McKegney, 1980) is insufficient, because they on 1.y assess the physical-functional level of the patients. Therefore, they do not cover all relevant aspects of quality oflife. In modern behavioural medicine, a multidimensional concept of quality of life is used (Aaronson, 1988; Fletcher, Hunt and Bulpitt, 1987; Hütter, 1990; Van Dam and Aaronson, 1988; Ware, 1984; Würtemberger, Hütter, Hirsch and Matthys, 1992). The Sickness Impact Profile (SIP), as a self-report measure of quality of life, was developed according to a psychosocial concept of quality of life which is assessed on a concrete behaviourallevel. It has proven reliability and validity Address forcorrespondence: B. 0. Hütter, Ph. D. Department ofNeurosurgery, Technical University, (RWTH) Aachen, 52057 Aachen, Germany, Pauwelsstr. 30. Phone 0241 /8088480; Fax 0241 /8888420.

149

•

150

B. O. HÜTTER AND G. WÜRTEMBERGER

(Bergner, Bobbitt, Pollard, Martin and Gilson, 1976a,b; Bergner, Bobbitt, Carter and Gilson, 1981; Follick, Smith and Ahern, 1985; Jones, Baveystock and Littlejohns, 1989; McSweeny, Grant, Heaton, Prigatano and Adams , 1985). In the present study, the internal consistency and split-half reliability of the German version of the SIP (Hütter, 1990) were investigated in patients with COPD and hypoxemia. It was the aim to corroborate the convergent validity of the German version of the SIP by means of studying the associations with different clinical and physiological indices of lung function and relevant subscales of the German personality form Freiburger Personality Inventory (FPI-R).

PATIENTS AND METHODS From January 1991 to June 1991, a total of 132 patients with COPD and hypoxemia were consecut{yely treated with liquid oxygen therapy (LOX) in the Department of Pneumonology of the University of Freiburg. The patients suffered from severe end stage COPD fulfilling the NOTT criteria (Benson... 1987; Nocturnal Oxygen Therapy Trial group, 1980). From May to June 1991, all patients were invited by mail to a medical follow-up examination. A set of questionnaires including the German version of the Sickness Impact Profile (SIP) was added to the letter. The subjects were·asked to fill out the questionnaires at horne and to send them back to the· Department of Pneumonology. Out'of the 132 patients, 8 (6%) had died in the meantime and 15 (11 %) had moved to an unknown address. A further 11 (8%) patients had been institutionalized because of hypoxic dementia and/or delirium. Of the remaining 98 patients who could be contacted, 63 (64%) were willing to fill out the questionnaire and to take part in the pneumonological follow -up examination. Five of them had to be excluded because they omitted several items of the SIP. Therefore, the present study includes aseries of 58 hypoxemic patients with COPD. At the time of the study (May/June 1991) all patients had already received LOX for one to five months. Table 1 gives the pneumonological and functional data of the patients in the study group.

Table 1

Physiological and c1inicaJ characteristics of the patients in the study sampie (n = 58)

Age (yrs)

Mean

SD

Min

Max

54.4

21.4

15.0

76.0

0.6

2.4

FVC (% pred)

29.3

8.5

15.0

45.0

Pa0 2 (mrnHg)

56.5

5.6

48.0

64.0

Sa0 2 (mrnHg)

84.6

4.2

76.0

92.0

FEV, (L/ s)

1.6#

Impairment (BDI)'

2.0#

0.0

3.0

Effort (BDI) '

2.0#

0.0

3.0

Summary score (BDI)'

4.0#

0.0

6.0

# Median

'BDI = Baseline Dyspnea Index (Mahler

er ai.

1984)

CHRONIC OBSTRUCTIVE PULMONARY DISEASE

151

A total of 47 (80 %) COPD patients were retired, 37 (64%) of whom as a consequence of their illness. Out of the 11 (20%) patients who worked regularly , eight did only light work and seven reported that they could only work for a short time and had to rest frequently. Of the 58 participants, 21 (37 %) had been treated before with common oxygen concentrators. The mean age was 54.4 years with a range from 15 to 76 years. The 15 year-old patient was a girl with severe cystic fibrosis . There were 49 (84 %) male and nine (16%) female participants. In September 1992, a survey of fatalities was performed in order to compute the one-year survival rate. From May 1991 to September 1992, 14 (24 %) patients died of their illness. This number is somewhat higher than the overall mortality of oxygen-dependent COPD patients with a five-year survival rate of about 30% (Benson, 1987). However, the study participants suffered from severe end-stage COPD. Rating

0/ pneumonological /unctional capacity

The functional impairment and the magnitude ' of effort subscales of the Baseline Dyspnea Index (BDI) were employed to assess dyspnea and exercise capacity (Mahler, Harver, Rosiello and Daubenspeck, 1989). Validity and reliability of t~eBDI in patients with chronic respiratory disease have been demonstrated hy previous studies (Mahler et al., 1989; Mahler, Rosiello, Harver, Lentine, McGovern and Daubenspeck, 1987; Mahler, Weinberg, Wells and Feinstein, 1984). The ratings of each BDI functional impairment and BDI magnitude of effort score were added to complete a BDI summary score. The functional assessment was performed during the pneumonological follow-up examination by the second author who had at this time no information about the quality of life data of the patients. Assessment

0/ physiological parameters 0/ lung function

The physiological parameters of lung function were assessed during the pneumonological follow-up examination by physicians who were blinded to the quality of life data of the patients. For organisational reasons , the time-delay between the completion Qf the questionnaires and the assessment of lung function ranged between so me days and four weeks. The severity of airway obstruction was quantified by the forced expiratory volume in one second (FEV I) and the forced vital capacity (FVC). Both parameters were measured in the sitting position using a portable spirometer (Spirometrics model 2451). The FVC was expressed as a percentage of the predicted value. Predicted normal values were adapted from Morris , Koski and Johnson (1971). The Pa0 2 and the Sa02 were analyzed at the same time while the patients were breathing room air. The Pa0 2 , arterial carbon dioxide tension (PaCO) and the pH were determined using an AVL 995 blood gas analyzer, allowing the calculation of the Sa0 2 (Würtemberger, Müller, Matthys and Sokolov, 1994). Assessment

0/ quality o/life

For assessing quality of life, the German version of the Sickness Impact Profile (SIP) was employed (Hütter, 1990). The American version of the SIP is a self-report questionnaire that measures sickness-related complaints in 12 aspects of quality of life: sleep and rest; mobility; horne management; social interaction; ambulation; employment;

152

B. O. HÜTIER AND G. WÜRTEMBERGER

recreation and pastimes; eating; communication; emotional and affective behaviour; body care and movement; alertness (Bergner et al. , 1976a,b; Bergner et al. , 1981). It contains 136 items and takes about 20 minutes to complete. Since the items of the SIP are formulated on a concrete behavioural level, the SIP is an objective and verifiable measure of quality of life. In several studies a good convergent and divergent validity of the SIP has been shown (Bergner et al. , 1976b; Bergner et al. , 1981; Follick et al. , 1985; Jones et al., 1989; McSweeny et al. , 1982). The 24-hour retest-reliability is about .90 for the entire instrument, while the internal consistency is about .80 (Bergner et al. , 1976a). The SIP was translated into German, item for item, using a consensus procedure by the authors in dose collaboration with a coworker of the first author, who speaks North American English, but who has lived now in Germany for several years. In case of divergent translations the dis agreement was cleared. In the German translation of the SIP, the "Emotional and affective behaviour", the "Alertness", and the "Body care and movement" subscales were discarded because they showed insufficient reliability and/or validity in an unpublished pilot study (Hütter, 1990). Therefore, the German version of the SIP contains 111 items in nine subscales, which enables the calculation of· nine subscale iffipairment scores, a sum score of total impairment, a psychosocial score including the "Social interaction", "Recreation and pastimes", and "Communication" subscales, and a physical score computed of the "Sleep and rest", "Ambulation", "Mobility", and "Horne-management" scales. In the instructions for each subscale of the SIP the respondents were requested to answer the questions as to how they perceived their symptoms on the day of filling out the form and further to report only such complaints which are related, according to their own opinion, to their present physical illness. Assessment of emotional adaptation and personality traits

For assessing the emotional state of the COPD patients, the FPI-R, a German standard personality form was employed (Fahrenberg, Hampel and Selg, 1984). It consists of 12 different bipolar subscales encompassing 138 items with dichotomous response possibilities (yes/no). The subscales of the FPI-R consist of 11 to 14 items and some items contribute to more than one subscale. The transformation of the raw scores of every subscale into stanine scores was adjusted for age and sex (Fahrenberg, Hampel and Selg, 1984). The FPI-R incorporates personality traits as well as personality disturbances typical for people with somatic complaints. The dimensions of the FPI Rare 1. life-satisfaction vs. life-dissatisfaction, 2. social responsibility vs . selfishness, 3. ambition vs. loss of motivation, 4. social insecurity vs. assertiveness, 5. sensitivity vs. calm, 6. aggressiveness vs. inaggressiveness, 7. strain vs. stability, 8. bodily complaints vs. health, 9. bodily concern vs. no bodily concern, 10. openness vs. social desirability, 11. extraversion vs. introversion and 12. emotionallability vs. emotional stability. To control for minimizing or social desirability response bias confounding the life quality assessment by the SIP, the social desirability subscale of the FPI-R was employed. Empirical work has demonstrated that this subscale corresponds to such response sets as social desirability and disclosedness (Amelang and Borkenau, 1981). In several studies adequate validity and reliability of the FPI-R subscales could be demonstrated (Fahrenberg et al. , 1984).

I

Psychology and Health, 1997, Vol. 12, pp. 149-159 Reprints available directly from the Publisher Photocopying permitted by license only

© 1997 OPA (Overseas Publishers Association) Amsterdam B. V. Published in The Netherlands by Harwood Academic Publishers Printed in Malaysia




PULMONARY DISEASE

B. O. HÜTIER 1 and G. WÜRTEMBERGER2 lDepartment 01 Neurosurgery, Technical University (RWTH) Aachen, FRG 2III. Medizinische Klinik, Carl-Thiem Klinikum, Cottbus , FRG (Received 17 August, 1994; in final form 4 Oerober, 1995)

Reliability and validity of the German version of the Sickness Impact Profile (SIP) were investigated in a sampie of 58 O, -dependent patients with chronic obstructive pulmonary disease (COPD). Internal consistency (Cronbach Alpha) was .93 for the entire instrument, while it was .85 far the SIP psycho-social scale and .89 for the SIP physical scale. The Cronbach Alphas ranged from .55 to .86 for the subscales of the German SI? version. The subscale HEating" was discarded because of low internal consistency (.40) and split-half reliability (.24). Indices ofrespiralOry function (FEVI' FVC, PaO" SaO,), functional impainnent and dyspn~a (BDI) were substantially associated with the total SIP, psychosocial and physical score, ranging from r = - .33 to r = -.48. A stepwise discriminant analysis between the 14 patients, who died during the course of the study and the survivors revealed one single discriminant function with an Eigenvalue of .5859 (canonical correlation r = .61; Chi' = 21.21; df =3; p < 0.001). The FVC (Wilk ' s Lambda .76), the severity of dyspnea (BOI) (Wilk's Lambda .71 ), the SIP total score (Wilk ' s Lambda of .65) and the FEV 1 (Wilk 's Lambda .63) were identified as discriminating variables, allowing the correct class ification of 85 % of the survivors and 91 % of the fatalitie s. Keywords: Quality of life, Sickness Impact Proft.le, German version of the Sickness Impact Profile, Psychometric analysis, Chronic obstructive pulmonary disease, Liquid oxygen therapy.

INTRODUCTION In modern medicine, quality of life has become an important criterion for assessing disease impact and treatment outcome (Aaronson, 1988; Van Dam and Aaronson, 1988; Hütter, 1990; Najman and Levine, 1981; Ware, 1984). For assessing quality of life, the use of current rating scales like the Glasgow Outcome Scale (Hütter and Gilsbach, 1993; Jennett and Bond, 1975; Jennett, Snoek, Bond and Brooks, 1981) or the Karnofsky Perfonnance Status Scale (Hutchinson, Boyd, Feinstein, Gonda, Hollomby and Rowat, 1979; Yates , Chalmer and McKegney, 1980) is insufficient, because they on ly assess the physical-functionallevel of the patients. Therefore, they do not cover aU relevant aspects of quality of life. In modern behavioural medicine, a multidirnensional concept of quality of life is used (Aaronson, 1988; Fletcher, Hunt and Bulpitt, 1987; Hütter, 1990; Van Dam and Aaronson , 1988; Ware, 1984; Würtemberger, Hütter, Hirsch and Matthys, 1992). The Sickness Impact Profile (SIP), as a self-report measure of quality of life, was developed according to a psychosocial concept of quality of life which is assessed on a concrete behaviourallevel. It has proven reliability and validity Address for correspondence: B. 0. Hütter, Ph. D. Department ofNeurosurgery, Technical Urtiversity, (RWTH) Aachen, 52057 Aachen, Germany, Pauwelsstr. 30. Phone 0241/8088480; Fax 0241 /8888420.

149

•

150

B. 0. HÜTTER AND G. WÜRTEMBERGER

(Bergner, Bobbitt, Pollard, Martin and Gilson, 1976a,b; Bergner, Bobbitt, Carter and Gilson , 1981; Follick, Smith and Ahem, 1985; Jones , Baveystock and Littlejohns, 1989; McSweeny, Grant, Heaton, Prigatano and Adams, 1985). In the present study, the intemal consistency and split-half reliability of the German version of the SIP (Hütter, 1990) were investigated in patients with COPD and hypoxemia. It was the aim to corroborate the convergent validity of the German version of the SIP by means of studying the associations with different clinical and physiological indices of lung function and relevant subscales of the German personality form Freiburger Personality Inventory (FPI-R) .

PATIENTS AND METHODS From January 1991 to June 1991, a total of 132 patients with COPD and hypoxemia were consecut~yely treated with liquid oxygen therapy (LOX) in the Department of Pneumonology of the University of Freiburg. The patients suffered from severe end stage COPD fulfilling the NOTT criteria (Bensoil',. 1987; Noctumal Oxygen Therapy Trial group, 1980). From May to June 1991 , all patients were invited by mai! to a medical follow-up examination. A set of questionnaires including the German version of the Sickness Impact Profile (SIP) was added to the letter. The subjects were 'asked to fill out the questionnaires at horne and to send them back to the Department of Pneumonology. Out'bf the 132 patients, 8 (6%) had died in the meantime and 15 (11 %) had moved to an unknown address. A further 11 (8%) patients had been institutionalized because of hypoxic dementia and/or delirium. Of the remaining 98 patients who could be contacted, 63 (64%) were willing to fill out the questionnaire and to take part in the pneumonological follow-up examination. Five of them had to be excluded because they omitted several items of the SIP. Therefore, the present study includes aseries of 58 hypoxemic patients with COPD. At the time of the study (May/June 1991) all patients had already received LOX for one to five months. Table 1 gives the pneumonological and functional data of the patients in the study group.

Table 1

Physiological and clinical characteristics of the patients in the study sampie (n = 58)

Age (yrs) FEV, (1../s)

Mean

SD

Min

Max

54.4

21.4

15.0

76.0

0.6

2.4

1.6#

FVC (% pred)

29.3

8.5

15 .0

45.0

Pa0 2 (mrnHg)

56.5

5.6

48.0

64.0

Sa0 2 (mmHg)

84.6

4.2

76.0

92.0

Impairrnent (BDI)'

2.0#

0.0

3.0

Effor! (BDI)'

2.0#

0.0

3.0

Summary score (BDI)'

4.0#

0.0

6.0

# Median

'BOl = Baseline Dyspnea Index (Mahler

€I a/., 1984)

CHRONIC OBSTRUCTlVE PULMONARY DISEASE

151

A total of 47 (80%) COPD patients were retired, 37 (64%) of whom as a consequence of their illness. Out of the 11 (20%) patients who worked regularly, eight did only light work and seven reported that they could only work for a short time and had to rest frequently. Of the 58 participants, 21 (37%) had been treated before with common oxygen concentrators. The mean age was 54.4 years with a range from 15 to 76 years. The 15 year-old patient was a girl with severe cystic fibrosis. There were 49 (84%) male and nine (16%) female participants. In September 1992, a survey of fatalities was performed in order to compute the one-year survival rate. From May 1991 to September 1992, 14 (24%) patients died of their illness. This number is somewhat higher than the overall mortality of oxygen-dependent CO PD patients with a five-year survival rate of about 30% (Benson, 1987). However, the study participants suffered from severe end-stage COPD.

Rating 01 pneumonological lunctional capacity The functional impairment and the magnitude ' of effort subscales of the Baseline Dyspnea Index (BDI) were employed to assess dyspnea and exercise capacity (Mahler, Harver, Rosiello and Daubenspeck, 1989). Validity and reliability of t~e BDI in patients with chronic respiratory disease have been demonstrated hy previous studies (Mahler et al. , 1989; Mahler, Rosiello , Harver, Lentine, McGovern and Daubenspeck, 1987; Mahler, Weinberg, Wells and Feinstein, 1984). The ratings of each BDI functional impairment and BDI magnitude of effort score were added to complete a BDI summary score. The functional assessment was performed during the pneumonological follow-up examination by the second author who had at this time no information about the quality of life data of the patients.

Assessment 01 physiological parameters 01 lung lunction The physiological parameters of lung function were assessed during the pneumonological follow-up examination by physicians who were blinded to the quality of life data of the patients. For organisational reasons, the time-delay between the completion Qf the questionnaires and the assessment of lung function ranged between some days and four weeks. The severity of airway obstruction was quantified by the forced expiratory volume in one second (FEV I) and the forced vital capacity (FVC). Both parameters were measured in the sitting position using a portable spirometer (Spirometrics model 2451) . The FVC was expressed as a percentage of the predicted value. Predicted normal values were adapted from Morris, Koski and Johnson (1971). The Pa0 2 and the Sa02 were analyzed at the same time while the patients were breathing room air. The Pa0 2 , arterial carbon dioxide tension (PaCO) and the pH were determined using an A VL 995 blood gas analyzer, allowing the calculation of the Sa0 2 (Würtemberger, Müller, Matthys and Sokolov, 1994).

Assessment 01 quality 01 life For assessing quality of life, the German version of the Sickness Impact Profile (SIP) was employed (Hütter, 1990). The American version of the SIP is a self-report questionnaire that measures sickness-related complaints in 12 aspects of quality of life: sleep and rest; mobility; horne management; social interaction; ambulation; employment;

152

B. O. HÜlTER AND G. WÜRTEMBERGER

recreation and pastimes; eating; communication; emotional and affective behaviour; body care and movement; alertness (Bergner et al., 1976a,b; Bergner et al., 1981 ). It contains 136 items and takes about 20 minutes to complete. Since the items of the SIP are formulated on a concrete behavioural level, the SIP is an objective and verifiable measure of quality of life. In several studies a good convergent and divergent validity of the SIP has been shown (Bergner et al., 1976b; Bergner et al., 1981 ; Follick et al., 1985; Jones et al., 1989; McSweeny et al., 1982). The 24-hour retest-reliability is about .90 for the entire instrument, while the internal consistency is about .80 (Bergner et al., 1976a). The SIP was translated into German, item for item, using a consensus procedure by the authors in elose collaboration with a coworker of the first author, who speaks North American English, but who has lived now in Germany for several years. In case of divergent translations the disagreement was eleared. In the German translation of the SIP, the "Emotional and affective behaviour", the "Alertness", and the "Body care and movement" subscales were discarded because they showed insufficient reliability and/or validity in an unpublished pilot study (Hütter, 1990). Therefore, the German version of the SIP contains 111 items in nine subscales, which enables the calculation of · nine subscale in\pairment scores, a sum score of total impairment, a psychosocial score ineluding the "Social interaction", "Recreation and pastimes", and "Communication" subscales, and a physical score computed of the "Sleep and rest", "Ambulation", "Mobility", and "Horne-management" scales. In the instructions for each subscale of the SIP the respondents were requested to ans wer the questions as to how they perceived their symptoms on the day of filling out the form and further to report only such complaints which are related, according to their own opinion, to their present physical illness. Assessment

0/ emotional adaptation and personality traits

For assessing the emotional state of the COPD patients, the FPI-R, a German standard personality form was employed (Fahrenberg, Hampel and Selg, 1984). It consists of 12 different bipolar subscales encompassing 138 items with dichotomous response possibilities (yes/no). The subscales of the FPI-R consist of 11 to 14 items and some items contribute to more than one subscale. The transformation of the raw scores of every subsca1e into stanine scores was adjusted for age and sex (Fahrenberg, Hampel and Selg, 1984). The FPI-R incorporates personality traits as well as personality disturbances typical for people with somatic complaints. The dimensions of the FPI Rare 1. life-satisfaction vs. life-dissatisfaction, 2. social responsibility vs. selfishness, 3. ambition vs. loss of motivation, 4. social insecurity vs. assertiveness, 5. sensitivity vs. calm, 6. aggressiveness vs. inaggressiveness, 7. strain vs. stability, 8. bodily complaints vs. health, 9. bodily concern vs. no bodily concern, 10. openness vs. social desirability, 11. extraversion vs. introversion and 12. emotionallability vs. emotional stability. To control for minimizing or social desirability response bias confounding the life quality assessment by the SIP, the social desirability subscale of the FPI-R was employed. Empirical work has demonstrated that this subscale corresponds to such response sets as social desirability and diselosedness (Amelang and Borkenau, 1981). In several studies adequate validity and reliability of the FPI-R subscales could be demonstrated (Fahren berg et al., 1984).

I