The side-effects to CPAP treatment inventory: the development and initial validation of a new tool for the measurement of side-effects to CPAP treatment

The side-effects to CPAP treatment inventory: the development and initial validation of a new tool for the measurement of side-effects to CPAP treatment
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  doi: 10.1111/j.1365-2869.2010.00825.x The side-effects to CPAP treatment inventory: the developmentand initial validation of a new tool for the measurement of side-effects to CPAP treatment ANDERS BROSTRO ¨ M 1 , 2 , KRISTOFER FRANZE ´ N A ˚RESTEDT 3 , 4 ,PER NILSEN 5 , ANNA STRO ¨ MBERG 3 , 6 , MARTIN ULANDER 2 , 7 and EVA SVANBORG 2 , 7 1 Department of Nursing Science, School of Health Sciences, Jo ¨nko ¨ping University, Jo ¨nko ¨ping,  2 Department of Clinical Neurophysiology,University Hospital, Linko ¨ping,  3 Division of Nursing, Department of Medical and Health Sciences, Linko ¨ping University, Linko ¨ping,  4 School of Human Sciences, University of Kalmar, Kalmar,  5 Division of Health and Society, Department of Medical and Health Sciences, Linko ¨pingUniversity,  6 Department of Cardiology, University Hospital and  7 Faculty of Health Sciences, Institution of Clinical and Experimental Medicine,Linko ¨ping University, Linko ¨ping, SwedenAccepted in revised form 04 November 2009; received 25 June 2009 SUMMARY  Continuous positive airway pressure (CPAP) is the treatment of choice for obstructivesleep apnoea syndrome (OSAS), but side-effects are common. No validated self-ratingscale measuring side-effects to CPAP treatment exists today. The aim was to develop theside-effects to CPAP treatment inventory (SECI), and investigate the validity andreliability of the instrument among patients with OSAS. SECI was developed on thebasis of: (1) in-depth interviews with 23 patients; (2) examination of the scientificliterature and (3) consensus agreement of a multi-professional expert panel. This yielded15 different types of side-effects related to CPAP treatment. Each side-effect has threesub-questions (scales): perceived frequency (a) and magnitude (b) of the side-effect, aswell as its perceived impact on CPAP use (c). A cross-sectional descriptive design wasused. A total of 329 patients with OSAS with an average use of CPAP treatment for39 months (2 weeks to 182 months) were recruited. Data were collected with SECI, andobtained from medical records (clinical variables and data related to CPAP treatment).Construct validity was confirmed with factor analysis (principal component analysiswith orthogonal rotation). A logical two-factor solution, the device subscale andsymptom subscale, emerged across all three scales. The symptom subscale describingphysical and psychological side-effects and the device subscale described mask anddevice-related side-effects. Internal consistency reliability of the three scales was good(Cronbach  s  a  = 0.74–0.86) and acceptable for the subscales (Cronbach  s  a  = 0.62– 0.86). The satisfactory measurement properties of this new instrument are promising andindicate that SECI can be used to measure side-effects to CPAP treatment. keywords  adherence, continuous positive airway pressure, obstructive sleep apnoeasyndrome, reliability, side-effects, validity INTRODUCTION Continuous positive airway pressure (CPAP) is currently thetreatment of choice in most cases of obstructive sleep apnoeasyndrome (OSAS; American Academy of Sleep Medicine TaskForce, 1999; Young  et al. , 2002). Correctly, administered to amotivated and adherent patient, CPAP can normalize noctur-nal breathing and reduce daytime symptoms (Giles  et al. ,2006). Long-term CPAP treatment can also lower theincreased morbidity and mortality attributable to OSAS(Marin  et al. , 2005). While this treatment is effective in theory, Correspondence : Anders Brostro ¨m, Department of Neurophysiology,University Hospital, S-581 85 Linko ¨ping, Sweden. Tel.: 13 222534; fax:13 224541; e-mail:  J. Sleep Res.  (2010) 19, 603–611  Assessing side-effects of CPAP   2010 European Sleep Research Society  603  several side-effects occur (Brostro ¨m  et al. , 2009a,b), and long-term adherence tends to be low (Haniffa  et al. , 2004; Weaver,2006). Adherence has been defined by the World HealthOrganization (WHO) as   the extent to which a person  sbehaviour taking medications, following a diet and  ⁄   or exe-cuting lifestyle changes, corresponds with agreed recommen-dations from a healthcare provider   (Sabate, 2003). Earlierstudies have estimated adherence to CPAP treatment to be inthe range of 46–80%, and the initial refusal to engage intreatment in the range of 8–15% (Engleman and Wild, 2003).In general, adherence to pharmacological treatment hasbeen found to be related to patient, disease and treatmentcharacteristics (Sabate, 2003). However, determinants forCPAP use have been difficult to identify. Greater diseaseseverity verified by objective sleep registration [Apnoea Hyp-opnoea Index (AHI)  ‡  15 and lower minO 2 ] and   ⁄   or highersleepiness [Epworth Sleepiness Scale (ESS) score  ‡  10] and  ⁄   orseverity of symptoms have shown weak correlations toincreased CPAP use (Engleman and Wild, 2003; Haniffa  et al. ,2004; Lewis  et al. , 2004; Weaver, 2006). However, it had to benoted that no universally agreed-upon definition exists of whatconstitutes CPAP treatment adherence, as the requiredamount of time on CPAP needed to obtain health benefits isnot known (Giles  et al. , 2006). A cut-off level of at least 4 h useper night (i.e. at least half of the expected sleep time) iscommonly applied as a criterion for regular CPAP use (Kribbs et al. , 1993).Both older (D  Ambrosio  et al. , 1999; Engleman  et al. , 1996;Hoffstein  et al. , 1992; Hui  et al. , 2001; Janson  et al. , 2000;Kalan  et al. , 1999; Pe ´pin  et al. , 1995) and more recent(Brostro ¨m  et al. , 2007, 2009a,b; Lindberg  et al. , 2006; Meurice et al. , 2007; Richards  et al. , 2007) studies have found thatblocked up nose, pressure sores from the mask, mask leaks,and dry mouth are common side-effects to CPAP treatment.Adherence might be reduced even further by factors such asavoidance of wearing the mask in front of others or turning off the machine because of disturbing noise (Brostro ¨m  et al. , 2008,2009a,b). Some of these side-effects, for example blocked upnose and dry mouth, might be alleviated using technicalsolutions, such as air humidifiers (Haniffa  et al. , 2004; Mador et al. , 2005). Psychological and patient-related factors can alsobe of importance (Stepnowsky  et al. , 2002; Tyrrell  et al. ,2006). For example, Type D personality (i.e. a combination of social inhibition and a tendency to experience negativeemotions) has been associated with an increase in perceivedfrequency and magnitude of CPAP-related side-effects, as wellas decreased adherence (Brostro ¨m  et al. , 2007).CPAP-treated patients with OSAS have been found to havepartially different views and perceptions from healthcarepersonnel regarding frequency of side-effects, magnitude of the problems and the problems   impact on CPAP use(Brostro ¨m  et al. , 2009a,b). Assessment of side-effects to CPAPtreatment can help to develop more tailored strategies aimed ateliminating or ameliorating problems that are likely to impactpatient adherence negatively (Brostro ¨m  et al. , 2009a,b). How-ever, although several studies have measured adherence, thereare currently no validated and tested instruments for side-effects to CPAP treatment. This study aims to fill an importantknowledge gap by describing the development and initialtesting of a new instrument, the side-effects to CPAP treatmentinventory (SECI), to measure the frequency and magnitude of side-effects, as well as their impact on CPAP use. Theinstrument  s validity and reliability among CPAP-treatedpatients with OSAS has been analysed. MATERIALS AND METHODSDevelopment and description of the SECI To establish content validity, the first step of the developmentof SECI consisted of 23 in-depth interviews with CPAP usersto obtain knowledge about experienced side-effects (Brostro ¨m et al. , 2009a,b). In parallel, we examined the scientific litera-ture in order to identify documented side-effects. A tentativelist of 20 side-effects was assembled on the basis of theinterviews and literature review. Next, the list of side-effectswas scrutinized and discussed by a multi-professional expertpanel consisting of three physicians specialized in sleepmedicine, three nurses working primarily with CPAP treat-ment, as well as two nurse researchers with experience of OSAS. Consensus among the panel members resulted in aconclusive list of 15 different types of side-effects. To reflect theclinical complexity, we decided that each side-effect had to beconsidered in three dimensions (referred to as scales from hereon): frequency of the side-effect (Frequency scale); magnitudeof the side-effect (Magnitude scale); and the impact of the side-effect on CPAP use (Impact scale). This resulted in a total of 45items. A five-point Likert-type scale was deemed appropriatefor each item. Treating the three scales separately, the possiblerange for each scale was 15–75. A higher score indicates ahigher frequency of side-effects, higher magnitude of side-effects and stronger impact on CPAP use. To strengthencontent and face validity during the development phase, agroup consisting of 20 patients with OSAS pilot-tested thisversion regarding readability, clarity and layout, with satisfy-ing results.The SECI was first developed in the Swedish language by themain author (A. Brostro ¨m). To secure equivalence whentranslating the SECI into English, an external bilingual groupexamined and approved the conceptual structure of theSwedish text. Then two of the authors (A. Brostro ¨m and M.Ulander) translated the scale into English. The Englishtranslation was examined by a bilingual group consisting of three sleep experts (physician, nurse and technician), as well asthree bilingual patients with OSAS, who proposed only a fewminor modifications of the wording. One of the authors (E.Svanborg), together with an external bilingual individual,translated the scale back into Swedish. Finally, the backtranslation of SECI was carefully examined by the externalbilingual group and decided to be equivalent to the srcinaltext. An example of the three scales for a side-effect taken fromthe SECI is presented in Table 1. 604  A. Brostro ¨ m  et al.   2010 European Sleep Research Society,  J. Sleep Res. ,  19 , 603–611  Design, setting and data collection for validation of the SECI A cross-sectional descriptive survey design was used to validatethe SECI. Patients were consecutively recruited from threeSwedish CPAP clinics (one university hospital, one provincialhospital and one private clinic) located in a metropolitan cityand two provincial cities in different parts of Sweden. Inclusioncriteria were age over 18 years and CPAP treatment for2 weeks or more based on diagnosis of OSAS (clinicalsymptoms and AHI > 10). Exclusion criteria were sufferingfrom another life-threatening disease, a diagnosis of a seriouspsychiatric disease, dementia, communication problems, orinability to read and speak Swedish. The CPAP initiation wasroutinely based on three 1-h visits at the CPAP clinic during a2-week period. During these visits, information about OSAS(e.g. causes of sleep apnoea and negative effects on sleep andhealth) and information about side-effects and practicalexercises related to the CPAP device were provided byhealthcare personnel.A questionnaire collecting data pertaining demographics,side-effects and adherence to CPAP treatment was provided to350 eligible patients with OSAS who fulfilled the inclusioncriteria. Clinical variables (co-morbidities, blood pressure,body mass index, ESS score, OSAS severity variables) andobjectively measured adherence to CPAP treatment (hours of CPAP use  ⁄   night) were obtained from medical records. Datafor the SECI were obtained on average 39 months (range2 weeks to 182 months) after patients   CPAP initiation. Theaverage time elapsed between the SECI data collection inrelation to collection of adherence data from the CPAP devicewas 4.5 months (range 0–8 months), and varied betweenpatients depending on the time elapsed since the last visit tothe clinic when adherence data were collected from the CPAPdevice. The study protocol was approved by the RegionalEthical Review Board at Linko ¨ping University, Sweden. Statistical processing and analysis The statistical methods used were dependent of the level of thedata, and both parametric and non-parametric statistics wereused (Altman, 1991). To validate the internal consistency of SECI, item analyses were performed for the three scales. Theseitem analyses included item-total correlations corrected foroverlaps andCronbach  salpha(Cronbach,1951)if itemdeleted(Nunnally and Bernstein, 1994). Spearman  s rank correlationwas used to investigate the relationship among items as well asamong the three scales (Fayers and Machin, 2007).A series of exploratory factor analyses (principal compo-nent, PCA) were performed to investigate the complexinterrelationships among the variables and the dimensionalityof the three scales in the SECI (i.e. the construct validity;Nunnally and Bernstein, 1994; Pett  et al. , 2003). Data werefirst examined with Bartlett  s test of sphericity, as well as withmeasure of sample adequacy in each variable (individualmeasures of sampling adequacy) and overall (Kaiser–MeyerOlkin measure). All these examinations indicated great sam-pling adequacy. All scales were analysed in which the numbersof factors extracted was decided by the Kaiser criteria(eigenvalue < 1.0). The scales were then analysed with twopredetermined factors, based on the Scree plot criteria. Anorthogonal (varimax) rotation was performed for all analyses.Factor loadings  ‡  0.4 were judged as significant according tothe sample size. The factor models were finally cross-validatedby a random split of the sample in two groups followed by newfactor analyses for comparisons against the whole sample(Hair  et al. , 1998).Known group validity test (i.e. a form of construct validity)was used to investigate if the SECI scales could discriminatebetween people with different adherence to CPAP treatment.Data on objective adherence to CPAP treatment were definedas the number of hours per night that the patients used thedevice. The CPAP use was dichotomized as machine usageabove or below 4 h mean use per night. The comparisonsbetween the two groups were then analysed using the Mann– Whitney  U  -test (Altman, 1991). The level of statistical signif-icance was overall set to  P  < 0.05. All statistical analyses wereperformed with the  spss ( spss  Inc, Chicago, IL, USA.). RESULTSStudy population A total of 329 patients out of 350 eligible patients with OSAS(94%) answered the questionnaire after one reminder. Patient Table 1  An example of a side-effect (Nasal stuffiness) revealed by three sub-questions focusing on the Frequency scale, the Magnitude scale andthe Impact scale taken from the SECINasal stuffinessHow frequently does this side-effectappear?Never 1 Seldom 2 Sometimes 3 Often 4 Very often 5How great a problem does thisside-effect cause?No problems 1 Small problems 2 Some problems 3 Great problems 4 Very great problems 5How does this side-effect decrease youruse of CPAP?Not at all 1 A little 2 Moderately 3 Much 4 Very much 5The other 14 side-effects included in the questionnaire are: runny nose, nose bleed, dry throat, irritated eyes, irritated bowl, transient deafness,feeling uncomfortable because of wearing CPAP in front of others, increased awakenings, uncomfortable pressure from the mask, mask leaks,cold air, disturbing noise, problems to exhale and anxiety during treatment.CPAP, continuous positive airway pressure. The side-effects to CPAP treatment inventory  605   2010 European Sleep Research Society,  J. Sleep Res. ,  19 , 603–611  characteristics are summarized in Table 2. All patients usedCPAP devices with fixed pressure and 21% of the patients useda humidifier. The number of months after CPAP initiation wason average 39 months (range 2 weeks to 182 months). Hyper-tension and diabetes were the most commonly reportedco-morbidities. The 21 non-responders did not differ fromthe responders regarding age, education level, marital status,co-morbidities, OSAS variables or length of time on CPAP. Validity of the SECI The item analysis demonstrated an overall satisfying item-totalcorrelation, albeit with some exceptions (Table 3). The itemsconcerning   nose bleed   demonstrated weak item-total correla-tions across the scales (0.13–0.18). The items   feeling uncom-fortable because of wearing CPAP   and   disturbing noise   alsohad item-total correlations  £  0.3, but only in the Frequencyscale. Except for the item   nose bleed  , the item-total correla-tions in general were weakest for the Frequency scale (range0.17–0.46), followed by the Magnitude scale (range 0.35–56)and the Impact scale (range 0.37–0.68). The median score forthe three scales varied between 22 and 31, with the highestscore for the Frequency scale and the lowest for the Impactscale. The strongest correlation between the three scales wasdemonstrated between the Frequency and Magnitude scale(0.87), whereas the weakest correlation was found between theFrequency and Impact scales (0.67; Table 3).The first set of factor analyses (not demonstrated in detail inthis study) was based on the Kaiser criteria. A five-factorsolution for the Frequency scale and a four-factor solution forthe Magnitude and Impact scale were identified. Aside fromthe different numbers of factors across the scales, a largeamount of items loaded on different factors, which made theinterpretation somewhat difficult. To determine the mostimportant factors the Scree plot criterion was used, whichindicated a possible two-factor structure. A second set of factor analyses with two predetermined factors was thereforeperformed.These factor analyses demonstrated a logical solution acrossthe scales (Table 4). Items related to the CPAP device such as  disturbing noise   and   mask leaks   loaded most strongly on onefactor across the scales. This dimension was therefore labelledthe Device subscale. Items related to symptoms of the CPAPtreatment, for instance   runny nose   and   irritated eyes  , loadedon the other factor. This was labelled the Symptom subscale.Even though the factor solutions were similar across the scales,a large amount of items ( n  = 8) in the Frequency scaledemonstrated communality values < 0.3. The Frequency scalealso demonstrated lower total explained variance (33%) thanthe Magnitude scale (40%) and the Impact scale (46%). Thecross-validation of the factor models (i.e. a random split of thesample in two groups followed by a new PCA) demonstratedthat the factor structure was fully reproduced for the Magni-tude scale and the Impact scale. However, the factor structurewas not reproduced for the Frequency scale as five itemsloaded on different factors compared with the factor analysison the whole sample.The known-group validity test showed that SECI was ableto discriminate between adherent and non-adherent CPAPusers (Table 5). Patients using CPAP  £  4 h per night scoredsignificantly higher median values (i.e. higher frequency of side-effects, increased magnitude of side-effects, and strongerimpact on CPAP use) for seven out of nine scales. TheFrequency  ⁄   Device subscale and the Magnitude  ⁄   Symptomsubscale were not able to discriminate between the two groupsat a significant level. Reliability of the SECI Most of the SECI scales demonstrated satisfactory internalconsistency reliability (Tables 3 and 4). Highest reliability wasseen in the Impact scale (0.87) and the Magnitude scale (0.83),whereas the Frequency scale had the lowest internal consis-tency (0.75; Table 3). The SECI subscales for Device andSymptoms related to the Magnitude and Impact scales showedsatisfactory internal consistency (0.72–0.86). The subscales for Table 2  Demographical and clinical data in CPAP-treated patientswith OSAS ( n  = 329)Gender:  n  (% within group)Men 263 (80)Women 66 (20)Age: mean years (SD, range) 60.3 years(9.3, 21–82)Educational level:  n  (% within group)Compulsory education (6 years) 94 (27)Further education (9 years) 74 (21)Higher education (12 years) 106 (30)University (15 years or more) 76 (22)Marital status:  n  (% within group)Married 234 (68)Unmarried 71 (20)Divorced 33 (9)Widow  ⁄   widower 12 (3)Self-reported co-morbidity:  n  (% within group)Hypertension 168 (51)Angina pectoris 60 (18)Acute myocardial infarction 30 (9)Stroke 13 (4)Diabetes 66 (20)BMI after CPAP: mean (range) 31.4 (20–49)OSAS variables: mean (range)AHI before CPAP 43.7 (10–92)ODI before CPAP 40.7 (10–90)Excessive daytime sleepiness: mean (SD)ESS before CPAP 12.1 (4.6)ESS with CPAP 6.9 (4.1)Time (months)After CPAP initiation: mean (range) 39 (2 weeks to182 months)CPAP pressure: mean (SD) 9.5 (2.4)AHI, apnoea–hypopnea index measured during sleep; BMI, bodymass index; CPAP, continuous positive airway pressure; ESS,Epworth Sleepiness Scale; ODI, oxygen desaturation index mea-sured during sleep; OSAS, obstructive sleep apnoea syndrome. 606  A. Brostro ¨ m  et al.   2010 European Sleep Research Society,  J. Sleep Res. ,  19 , 603–611  Device and Symptoms related to the Frequency scale hadlower internal consistency (0.62 and 0.67, respectively;Table 4). DISCUSSION Despite wide documentation of poor adherence to CPAPtreatment and prevalent side-effects (Haniffa  et al. , 2004;Weaver, 2006), there is to the best of our knowledge novalidated instrument to measure side-effects to CPAP. Wefound that SECI, which includes 15 different types of side-effects, showed satisfactory validity and reliability, suggestingthat the instrument can be used to identify patients whoexperience side-effects of CPAP and are at increased risk forlow adherence to CPAP treatment.Content validity (i.e. does the test measure all relevantcomponents of the property?) was established from 23 in-depthinterviews with CPAP-users, a literature review, as well as aconsensus agreement by a multi-professional expert panel.Results from the item analysis (Fayers and Machin, 2007;Nunnally and Bernstein, 1994) showed that the Magnitude andthe Impact scales showed good internal consistency andfunctioned as intended, whereas the Frequency scale demon-strated some weaknesses (Table 3). The Frequency scale as awhole had satisfactory internal consistency reliability, but thisscale had overall the lowest item-total correlations. Only fouritems had an item-total correlation > 0.4 in the Frequencyscale, compared with 12 items for the Magnitude scale and 13items for the Impact scale. However, none of the items in thelatter two scales showed correlations with other items > 0.7,indicating no redundancy of items. The Magnitude and theImpact scales demonstrated a high homogeneity with Cron-bach  s alpha values > 0.8. The high homogeneity is furthersupported by the fact that the alpha values of both scaleshardly changed if any of the items were deleted. If the internalconsistency of a scale is weak, some items might not measurewhatever the scale measures, whereas if internal consistency isextremely strong, it means that some items might be redun-dant. When developing the SECI, the frequency of a side-effect(i.e. The Frequency scale) appeared to be an important aspectto account for. However, it is possible that frequency of side-effects is less important than the perceived magnitude of the Table 3  Item analysis for the three scales in the SECI ( n  = 329) Frequency scale Magnitude scale Impact scaleMedian(q1–q3)Item-total correlation*Cronbach  s a  if itemdeleted Median(q1–q3)Item-total correlation*Cronbach  s a  if itemdeleted Median(q1–q3)Item-total correlation*Cronbach  s a  if itemdeleted  Item statisticsBlocked up nose 2 (2–3) 0.392 0.731 2 (2–3) 0.452 0.818 2 (1–3) 0.488 0.864Runny nose 2 (1–3) 0.447 0.726 2 (1–3) 0.485 0.815 1 (1–2) 0.506 0.862Nose bleed 1 (1–2) 0.178 0.749 1 (1–1) 0.160 0.832 1 (1–1) 0.131 0.878Dry throat 3 (2–4) 0.386 0.732 2 (2–3) 0.481 0.816 2 (1–3) 0.615 0.856Irritated eyes 2 (1–3) 0.396 0.731 2 (1–3) 0.516 0.813 1 (1–2) 0.522 0.861Irritated bowl 1 (1–2) 0.421 0.728 1 (1–2) 0.411 0.820 1 (1–1) 0.369 0.867Transient deafness 1 (1–2) 0.365 0.735 1 (1–1) 0.386 0.822 1 (1–1) 0.447 0.865Feeling uncomfortablebecause of wearingCPAP2 (1–3) 0.171 0.758 1 (1–2) 0.350 0.823 1 (1–1) 0.430 0.865Increased awakenings 2 (2–3) 0.394 0.731 2 (1–3) 0.501 0.814 1 (1–2) 0.673 0.853Uncomfortable pressureof the mask3 (2–3) 0.378 0.732 2 (2–3) 0.563 0.810 2 (1–2) 0.678 0.852Mask leaks 3 (2–3) 0.382 0.733 2 (2–3) 0.532 0.812 2 (1–3) 0.620 0.856Cold air 1 (1–2) 0.324 0.738 1 (1–2) 0.432 0.819 1 (1–1) 0.586 0.859Disturbing noise 2 (1–3) 0.259 0.747 2 (1–2) 0.424 0.820 1 (1–2) 0.564 0.859Problems of exhaling 1 (1–2) 0.459 0.726 1 (1–2) 0.504 0.814 1 (1–2) 0.605 0.857Anxiety duringtreatment1 (1–2) 0.402 0.732 1 (1–1) 0.455 0.818 1 (1–1) 0.528 0.861Scale statisticsMedian score (q1–q3) 31 (26–36) 26 (23–31) 22 (17–28)Cronbach  s  a  0.749 0.828 0.869Scale correlations  Frequency 1.000 0.874 0.674Magnitude 0.874 1.000 0.804Impact 0.674 0.804 1.000Possible score range: items 1–5; scales 15–75CPAP, continuous positive airway pressure.*Corrected for overlaps.  Spearman rank correlations. The side-effects to CPAP treatment inventory  607   2010 European Sleep Research Society,  J. Sleep Res. ,  19 , 603–611
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks