Qualitative studies evaluating quality of life (QoL) after coronary artery bypass surgery (CABS) suggest that many surgery specific factors interfere with patient function. Patients report incisional (sternotomy and donor graft leg) pain and drainage that persist 1 month after CABS, as well as continuous pain from the shoulders and neck, respiratory problems, feelings of weakness, and sleeping difficulties including chest wall pain with sidelying, waking frequently and early, and more nightmares than usual.1 Other important patient concerns have been identified as problems with wound healing, thoracic pain, and dissatisfaction with postoperative supportive care 1 year after CABS.2 Other studies have found discomfort at chest and leg incisions; shortness of breath; problems with eating; difficulty with sleeping; pain in the shoulders, back, and neck; ineffective coping; and medication side effects as frequently reported symptoms in patients after CABS.3,4
Functional status after CABS is often measured using generic, self-report QoL instruments.5-7 However, a potential disadvantage of generic health-related QoL instruments is an insensitivity to disease-specific health-related QoL parameters.6-8 Furthermore, cardiac-specific QoL instruments may rely on a concurrent diagnosis of angina (Seattle Angina Questionnaire) or heart failure (Minnesota Living with Heart Failure Questionnaire), which are not universally experienced by patients that have undergone CABS.9,10
Symptom inventories have been used to provide disease-specific information on patient-perceived QoL.3,11 The Heart Surgery Symptom Inventory (HSSI) was developed to provide a disease-specific outcome measure for patients after CABS. Using a standardized instrument to assess the impact of disease-specific symptoms would be useful for evaluating QoL, in addition to generic instruments, in patients after CABS. The purpose of this study was to evaluate the test-retest reliability, internal consistency, and concurrent validity of the HSSI in patients after CABS.
METHODS
Subjects
This study included 28 people (age = 66.6 +/- 10.5 years, 70% male) who had recently undergone CABS. Subjects were recruited from 2 outpatient cardiac rehabilitation programs. Criteria for study participation were less than 6 months post-CABS, age over 35 years, ability to understand the English language, and current participation in an outpatient cardiac rehabilitation program. Patients with significant cognitive deficits or who had sustained a cerebral vascular accident after surgery were excluded from study participation.
Procedures
Patients were given a packet of questionnaires to take home, complete, and return 2 days later at their next cardiac rehabilitation session. The participants were instructed to answer the questions in order, not to go back and review previous answers, and to finish the questionnaires in a single session. No time limit was specified but participants reported needing between 1 and 2 hours to complete all assessments. First, patients completed the appropriate subsections of the HSSI. Next, patients completed additional assessments of pain including a body chart and visual analog scale (VAS) and a generic 13-item symptom checklist. Next, patients completed the Medical Outcomes Study Short Form-36 (SF-36) and a second HSSI.
Instruments
The HSSI was developed and refined using data from previously published qualitative studies,1-4 subjective information from clinicians and patients recovering from CABS, and finally feedback from a panel of content experts. The content validity of the HSSI has been previously reported.12 The HSSI has 5 subcategories of symptoms: general, cardiac, trunk, lower extremity, and upper extremity. All patients completed the first 3 sections (general, cardiac, and trunk), and those who had undergone saphenous vein or radial artery harvesting also completed the lower extremity or upper extremity sections, respectively. The HSSI has a total of 76 items that are consistently worded negatively using the phrase, "During the past week, how much have you been bothered by[horizontal ellipsis]" This instrument uses a 5-point Likert scale with higher scores indicating greater severity of symptoms than lower scores.
The SF-36, assessments of pain (body chart and VAS), and a generic 13-item symptom checklist were used to assess concurrent validity. The SF-36 is a self-report instrument that measures generic health-related QoL and is widely considered the criterion measure of QoL. This instrument has been used extensively to study generic health-related QoL in patients with cardiac problems. In addition, measurements obtained with the SF-36 have well documented and acceptable degrees of reliability, validity, and sensitivity.6-8 Body charts and VAS are standard instruments to measure pain severity in a wide variety of patient populations. These instruments have acceptable degrees of reliability and validity in noncardiac populations.13,14 Generic symptom checklists are commonly used in clinical settings to obtain a standardized patient history. The symptom checklist used in this study should have similar reliability to other self-report instruments and strong face validity for assessing the presence or absence of symptoms in patients after CABS. Items in the generic symptom checklist included itching or irritation at surgical incisions, shortness of breath or difficulty breathing, sore throat, frequent cough, heart palpitations, dizziness or light-headedness, fatigue or weakness, stiffness or tightness, numbness or tingling, swelling, difficulty sleeping, difficulty looking at incisions, and difficulty eating.
Data Analyses
Items on the HSSI were summed to generate total and subcategory scores. The number of positive items on the generic symptom checklist was summed. Data regarding pain from the body chart were dichotomized to create nominal categories of pain or no pain. Length for the 10-cm VAS was measured manually. Results of the first and second HSSI were evaluated using Pearson correlations (r) and intraclass correlation coefficients (ICCs). Correlations among individual items, subcategory scores, and total score from the first trial of the HSSI were calculated. Scores on the HSSI and scores on the SF-36, VAS, and symptom checklist were evaluated using correlations. Degree of agreement between reports of pain on the body chart and VAS with HSSI was evaluated with chi-square. The [alpha] level was set at less than .05.
RESULTS
Test-Retest Reliability
Correlations between trials 1 and 2 for the HSSI total and all subcategories were significant. Correlations were total r = 0.98 and ICC = 0.56, cardiac r = 0.89 and ICC = 0.74, general r = 0.98 and ICC = 0.77, trunk r = 0.96 and ICC = 0.59, and lower extremity r = 0.94 and ICC = 0.75.
Internal Consistency
Individual item, subcategory, and total correlations were calculated using scores from the first trial of the HSSI. Subcategories to total correlations were all significant and ranged from 0.59 to 0.85. Subcategory to subcategory correlations were all significant except for lower extremity to cardiac and lower extremity to general comparisons (see Table 1). Individual items to total and subcategory correlations are shown in Table 2.
Concurrent Validity
Correlations between HSSI and SF-36 scores were calculated for 26 of the subjects participating in this study. Two subjects did not complete the SF-36. The relationships between the HSSI total score and all SF-36 scores were significant. The correlations for the HSSI total compared to SF-36 Physical and Mental Component Summary were -0.55 and -0.54, respectively. Correlations between the SF-36 and HSSI lower extremity were the lowest, and those between the SF-36 and HSSI general were the highest within the subcategory analysis. Correlations between the SF-36 and HSSI scores are shown in Table 3. The correlation between similar items on the HSSI and the symptom checklist was significant (r = 0.72). The correlation between pain items on the HSSI and the VAS was not significant. The percent of subjects reporting pain on the HSSI was 75%, which was not significantly different than the percent of subjects reporting the presence of pain on the body chart (71%) and VAS (79%).
DISCUSSION
The results of this study suggest that the HSSI has acceptable levels of reliability, internal consistency, and concurrent validity in patients recovering from CAB for clinical and research applications. Test-retest reliability is the degree of measurement consistency between multiple trials. Correlations greater than r = 0.90 are generally considered acceptable reliability for clinical and research applications.15 Therefore, the correlations between total and subcategory scores of the HSSI suggest acceptable reliability in patients after CABS. Internal consistency reflects the degree of agreement between items, subcategories, and total scores. Ideally, there should be some correlation between parts of an assessment, indicating that the items have some relationship to each other and should be grouped together, but not total agreement, indicating that the parts are redundant. Most of the correlations among HSSI items, subcategories, and total HSSI scores were between 0.25 and 0.75, indicating a fair to moderate relationship among components of the HSSI.15
Results of this study indicate that the HSSI has good concurrent validity with assessments of generic QoL and pain. Concurrent validity reflects the degree of agreement between measurements of similar phenomena obtained using different instruments. The strongest degree of relationship was found between total and general HSSI scores and physical function, role limitation due to physical health, and physical component summary of the SF-36. Scores on the lower extremity subcategory of the HSSI had the weakest relationship with scores on the SF-36. These findings suggest that total and most subcategory scores of the HSSI reflect generic QoL, but the lower extremity subcategory scores of the HSSI alone are not a good indicator of QoL. Despite the high incidence (>74%) of lower extremity symptoms among patients in this study, little relationship was found with indices of QoL. This finding may reflect that, unlike general, cardiac, and trunk symptoms, lower extremity symptoms do not alter patient activity patterns or ability and therefore do not negatively influence QoL.
Results on corresponding items of the HSSI showed good agreement with results on the generic symptom checklist and the presence of pain on the body chart and VAS. Surprisingly, there was no relationship between pain severity measured on the VAS and score of pain items on the HSSI. A possible explanation for this finding is that the VAS score reflects the overall severity of bodily pain which may occur at only 1 anatomic location, whereas the pain items on the HSSI reflect both prevalence and severity of pain which may occur at several anatomic locations.
Due to the design of this study, the results need to be interpreted with some caution. Because all testings took place in a single session lasting 1 to 2 hours, it is possible that the subjects remembered answers they provided on the first trial of the HSSI when completing the second trial of the HSSI. This effect may have been minimized because subjects were not allowed to review previous answers and because of the high number of responses. However, due to the dynamic nature of many symptoms, such as pain and shortness of breath, it is necessary to administer the 2 trials of the HSSI relatively close together. Test-retest reliability should be an indicator of the measurement instrument's stability and not the inherent variability of the phenomenon measured; therefore, close proximity of retesting in this study was warranted. In addition, all subjects were participants in an outpatient cardiac rehabilitation program which may have resulted in a biased sample; that is, it is not known whether patients participating in cardiac rehabilitation experience greater or fewer symptoms after CABS than those who are not participants. Lastly, subjects were less than 6 months post-CABS and therefore in the subacute stage of recovery (mean +/- SD = 99 +/- 39, range = 28-171 days since surgery). Use of the HSSI with patients during acute or long-term recovery from CABS should be done with caution because the instrument has not been evaluated for patients within those time periods.
A standardized instrument to assess the presence and impact of disease-specific symptoms will help to better elucidate the process of recovery from CABS. The HSSI may be useful in evaluating the influence of surgical techniques, rehabilitation, or premorbid factors on patient outcomes after CABS. In addition, the HSSI could assist rehabilitation professionals in identifying factors that contribute to patient functional limitations/disability, compliance with lifestyle changes, and risk for developing dependence with activities of daily living and instrumental activities of daily living. Understanding and minimizing surgery-related symptoms in patients recovering from CABS may help to hasten recovery, enhance QoL, and improve adherence with lifestyle modifications.
References