Structural equation modeling of the medical outcome study: Short Form 36 (SF-36)

The factorial validity of the SF-36 was evaluated using confirmatory factor analysis (CFA) methods, structural equation modeling (SEM), and multigroup structural equation modeling (MSEM). First, the measurement and structural model of the hypothesized SF-36 was explicated. Second, the model was tested for the validity of a second-order factorial structure, upon evidence of model misfit, determined the best-fitting model, and tested the validity of the best-fitting model on a second random sample from the same population. Third, the best-fitting model was tested for invariance of the factorial structure across race, age, and educational subgroups using MSEM.^ The findings support the second-order factorial structure of the SF-36 as proposed by Ware and Sherbourne (1992). However, the results suggest that: (a) Mental Health and Physical Health covary; (b) general mental health cross-loads onto Physical Health; (c) general health perception loads onto Mental Health instead of Physical Health; (d) many of the error terms are correlated; and (e) the physical function scale is not reliable across these two samples. This hierarchical factor pattern was replicated across both samples of health care workers, suggesting that the post hoc model fitting was not data specific. Subgroup analysis suggests that the physical function scale is not reliable across the "age" or "education" subgroups and that the general mental health scale path from Mental Health is not reliable across the "white/nonwhite" or "education" subgroups.^ The importance of this study is in the use of SEM and MSEM in evaluating sample data from the use of the SF-36. These methods are uniquely suited to the analysis of latent variable structures and are widely used in other fields. The use of latent variable models for self reported outcome measures has become widespread, and should now be applied to medical outcomes research. Invariance testing is superior to mean scores or summary scores when evaluating differences between groups. From a practical, as well as, psychometric perspective, it seems imperative that construct validity research related to the SF-36 establish whether this same hierarchical structure and invariance holds for other populations.^ This project is presented as three articles to be submitted for publication. ^

Health-related quality of life and survival in liver transplant candidates.

Health-related quality of life (HRQOL) is an important measure of the effects of chronic liver disease in affected patients that helps guide interventions to improve well-being. However, the relationship between HRQOL and survival in liver transplant candidates remains unclear. We examined whether the Physical Component Summary (PCS) and Mental Component Summary (MCS) scores from the Short Form 36 (SF-36) Health Survey were associated with survival in liver transplant candidates. We administered the SF-36 questionnaire (version 2.0) to patients in the Pulmonary Vascular Complications of Liver Disease study, a multicenter prospective cohort of patients evaluated for liver transplantation in 7 academic centers in the United States between 2003 and 2006. Cox proportional hazards models were used with death as the primary outcome and adjustment for liver transplantation as a time-varying covariate. The mean age of the 252 participants was 54 +/- 10 years, 64% were male, and 94% were white. During the 422 person years of follow-up, 147 patients (58%) were listed, 75 patients (30%) underwent transplantation, 49 patients (19%) died, and 3 patients were lost to follow-up. Lower baseline PCS scores were associated with an increased mortality rate despite adjustments for age, gender, Model for End-Stage Liver Disease score, and liver transplantation (P for the trend = 0.0001). The MCS score was not associated with mortality (P for the trend = 0.53). In conclusion, PCS significantly predicts survival in liver transplant candidates, and interventions directed toward improving the physical status may be helpful in improving outcomes in liver transplant candidates.

Dietary Intakes of Children From Food Insecure Households

The objective of this cross-sectional study is to compare the dietary behaviors of children from low food secure (LFS) households with children from very low food secure (VLFS) households over an entire day, and during meals specifically consumed at home —breakfast, snack, and dinner. Parents of the recruited children completed a demographic questionnaire, along with USDA's 6-item short form food security questionnaire. Children completed 24-hour dietary recalls in person. Complete data from 102 children 9 to 12 years old were used. Using ANOVA to assess the differences among groups in intakes over an entire day and during meals consumed at home, no significant differences were found, except for vitamin C intake during breakfast. Based on the definitions of LFS and VLFS, it was hypothesized that children from VLFS group would have lower dietary intakes compared to children from the LFS group. However, this study found little difference. Speculations have been made for these findings and implications for research have been presented.

The impact of antibiotic associated diarrhea on health related quality of life in hospitalized patients

Objective. To determine the impact of antibiotic associated diarrhea (AAD) on health related quality of life (HRQOL) in hospitalized patients compared to matched controls without diarrhea. ^ Methods. This is a hospital-based, matched case-control study using secondary data from a prospective cohort trial of patients receiving broad-spectrum antibiotics. One hundred and seventy-eight patients were recruited of whom 18 (10%) reported having antibiotic associated diarrhea. Two non-diarrhea controls were selected for each case with diarrhea giving a final sample of 18 cases and 36 controls. Responses from Short Form (SF) 36 questionnaire were aggregated into eight domains including physical functioning (PF), role-functioning physical (RP), bodily pain (BP), general health (GH), social functioning (SF), vitality (VT), role-functioning emotional (RE), and mental health (MH). The eight domains were compared between cases and controls. A GI targeted HRQOL measure was administered to 13 patients with AAD. Responses from the disease-specific instrument were combined in eight subscale scores: dysphoria, interference with activity, body image, health worry, food avoidance, social reaction, sex, and relationships. ^ Results. The sample consisted of 41 females (75.9%) and 13 males (24.1%) aged 53.5 ± 14.4 years (range: 21-76 years). Twenty five patients (46%) were Caucasian, 15 (27%) were African American, 13(24%) were Hispanic and 1(2%) was Asian. In univariate analysis, no significant differences in quality of life outcomes were observed in each of the SF36 domains between the case patients and matched controls. There were trends for decreased scores on the role-functioning physical, bodily pain, general health, social functioning, mental health, and mental summary domains. In total, 7 of 8 domain scores were lower in patients with AAD and 5 of 8 domain scores were lower by more than 5 points (considered clinically significant). Controlling for age, patients with antibiotic associated diarrhea had significantly lower general health, vitality, and mental health scale scores (p<0.05 each). The disease-specific scores were significantly lower in patients with AAD than those in published norms for irritable bowel syndrome patients. ^ Conclusion. In this small sample, several areas of decreased QOL in patients with AAD compared to matched controls were noted. A larger sample size to validate these results is necessary.^

Frequency, severity and distress of dialysis-related symptoms reported by patients on hemodialysis

Purpose. To determine which symptoms are the most reported, occur most frequently, have the greatest severity, and cause the most bother for hemodialysis (HD) patients and to determine if the symptoms experienced differ between the first (HD 1) and second (HD 2) treatments of the week. ^ Design. An observational, comparative design was used to determine participants' HD symptoms experience on HD 1 and HD 2, and the effect of the symptom experience on Quality of Life (QOL). One hundred subjects were recruited from five dialysis centers. ^ Methods. The adapted Dialysis Frequency, Severity and Symptom Burden Index (DFSSBI) and the Medical Outcomes Study Short Form 36 (MOS SF 36) were administered (N = 99) on HD 1 and the DFSSBI again on HD 2. Data were analyzed for significance among symptoms experience test scores in relation to HD 1 and HD 2, QOL, and gender and age. ^ Results. Of 31 symptoms assessed, respondents reported an average of 9.69 symptoms on HD 1 and 7.51 symptoms on HD 2. Overall, more symptoms were reported, and were more frequent, severe and bothersome on HD 1 when the level of metabolic waste is highest. The most reported symptoms included tiredness, dry skin, difficulty falling asleep, itching, numbness/tingling, difficulty staying asleep, decreased interest in sex, and bone/joint pain. Females scored consistently higher than males in the four symptom dimensions. The respondents reported about the same as the population norm (50) on the physical component summary score of the MOS SF 36 and higher than the norm (65.23) on the mental component summary score. ^ Conclusion. The study findings highlighted the fact that hemodialysis patients experience multiple symptoms that can be frequent, severe, and bothersome. Interventions should be developed and tested to reduce symptom burden and improve QOL. ^

THE MEASUREMENT OF HUMANISTIC AND DEHUMANISTIC BEHAVIORS IN HOSPITAL NURSING ENVIRONMENTS

The purpose of this investigation was to develop a reliable scale to measure the social environment of hospital nursing units according to the degree of humanistic and dehumanistic behaviors as perceived by nursing staff in hospitals. The study was based on a conceptual model proposed by Jan Howard, a sociologist. After reviewing the literature relevant to personalization of care, analyzing interviews with patients in various settings, and studying biological, psychological, and sociological frames of reference, Howard proposed the following necessary conditions for humanized health care. They were the dimensions of Irreplaceability, Holistic Selves, Freedom of Action, Status Equality, Shared Decision Making and Responsibility, Empathy, and Positive Affect.^ It was proposed that a scale composed of behaviors which reflected Howard's dimensions be developed within the framework of the social environment of nursing care units in hospitals. Nursing units were chosen because hospitals are traditionally organized around nursing care units and because patients spend the majority of their time in hospitals interacting with various levels of nursing personnel.^ Approximately 180 behaviors describing both patient and nursing staff behaviors which occur on nursing units were developed. Behaviors which were believed to be humanistic as well as dehumanistic were included. The items were classified under the dimensions of Howard's model by a purposively selected sample of 42 nurses representing a broad range of education, experience, and clinical areas. Those items with a high degree of agreement, at least 50%, were placed in the questionnaire. The questionnaire consisted of 169 items including six items from the Marlowe Crowne Social Desirability Scale (Short Form).^ The questionnaire, the Social Environment Scale, was distributed to the entire 7 to 3 shift nursing staff (603) of four hospitals including a public county specialty hospital, a public county general and acute hospital, a large university affiliated hospital with all services, and a small general community hospital. Staff were asked to report on a Likert type scale how often the listed behaviors occurred on their units. Three hundred and sixteen respondents (52% of the population) participated in the study.^ An item analysis was done in which each item was examined in relationship to its correlation to its own dimension total and to the totals of the other dimensions. As a result of this analysis, three dimensions, Positive Affect, Irreplaceability, and Freedom of Action were deleted from the scale. The final scale consisted of 70 items with 26 in Shared Decision Making and Responsibility, 25 in Holistic Selves, 12 in Status Equality, and seven in Empathy. The alpha coefficient was over .800 for all scales except Empathy which was .597.^ An analysis of variance by hospital was performed on the means of each dimension of the scale. There was a statistically significant difference between hospitals with a trend for the public hospitals to score lower on the scale than the university or community hospitals. That the scale scores should be lower in crowded, understaffed public hospitals was not unexpected and reflected that the scale had some discriminating ability. These differences were still observed after adjusting for the effect of Social Desirability.^ In summary, there is preliminary evidence based on this exploratory investigation that a reliable scale based on at least four dimensions from Howard's model could be developed to measure the concept of humanistic health care in hospital settings. ^