Performance tiers: Implementing comparative effectiveness analysis in the health care setting through operations research and spatial methods

This dissertation develops and tests a comparative effectiveness methodology utilizing a novel approach to the application of Data Envelopment Analysis (DEA) in health studies. The concept of performance tiers (PerT) is introduced as terminology to express a relative risk class for individuals within a peer group and the PerT calculation is implemented with operations research (DEA) and spatial algorithms. The analysis results in the discrimination of the individual data observations into a relative risk classification by the DEA-PerT methodology. The performance of two distance measures, kNN (k-nearest neighbor) and Mahalanobis, was subsequently tested to classify new entrants into the appropriate tier. The methods were applied to subject data for the 14 year old cohort in the Project HeartBeat! study.^ The concepts presented herein represent a paradigm shift in the potential for public health applications to identify and respond to individual health status. The resultant classification scheme provides descriptive, and potentially prescriptive, guidance to assess and implement treatments and strategies to improve the delivery and performance of health systems. ^

Identifying characteristics of mentally ill homeless individuals who are successfully linked to health and social services post-incarceration via the jail inreach project

Healthcare for the Homeless—Houston (HHH) received a research grant from The Medallion Foundation, Inc. in March 2006 to pilot The Jail Inreach Project, an intensive “inreach” initiative to assess the impact of providing continuity of mental and primary health care services for homeless individuals who suffer from mental illness and/or substance abuse being released from jail. This pilot project was initiated by HHH, in collaboration with the Harris County Sheriff’s Office and the Mental Health Mental Retardation Authority of Harris County (MHMRA). Those who are flagged as “frequent flyers” and who are diagnosed with a mental illness are referred to the Jail Inreach Project. In order to maximize the effectiveness of the discharge plan, case managers offer the option of meeting the client at the time of release and bring them to the HHH clinic located four blocks from the jail. Participation in both the program and the option for direct release to the care of a case manager are voluntary.^ The purpose of this study is to determine the outcomes of the Jail Inreach Project and addresses the following objectives: (1) to evaluate the characteristics of inmates that chose to be released from jail to the direct care of an HHH case manager versus those who opt for self release and (2) to determine the number and percent of inmates that are linked to services and relationship with type of release (direct versus indirect), (3) to determine if there is a relationship between outcomes and characteristics and (4) to determine what outcomes are a function of release, controlling for characteristics. Statistical analysis, including frequencies, cross tabulations, chi-square and logistical regression, found that those who opt for self release are six times less likely to be successfully linked to services and that gender is the most significant predictor of choosing self release. Men are far more likely to opt for self release than women engaged in this program. These findings help inform policy and program design and development that addresses the difference in service utilization and successful linkage to services post-incarceration. Successful linkage to services, thus continuity of and access to care, further impact the effects of the revolving door phenomenon of mentally ill homeless individuals cycling between the streets, jails and hospital emergency centers.^

Air Force Web Preventive Health Assessment (AF WEBPHA) mental health screening effectiveness

The U.S. Air Force assesses Active Duty Air Force (ADAF) health annually using the Air Force Web-based Preventative Health Assessment (AF WebPHA). The assessment is based on a self-administered survey used to determine the overall Air Force health and readiness, as well as, the individual health of each airman. Individual survey responses as well as groups of responses generate further computer generated assessment and result in a classification of 'Critical', 'Priority', or 'Routine', depending on the need and urgency for further evaluation by a health care provider. The importance of the 'Priority' and 'Critical' classifications is to provide timely intervention to prevent or limit unfavorable outcomes that may threaten an airman. Though the USAF has been transitioning from a paper form to the online WebPHA survey for the last three years it was not made mandatory for all airmen until 2009. The survey covers many health aspects including family history, tobacco use, exercise, alcohol use, and mental health. ^ Military stressors such as deployment, change of station, and the trauma of war can aggravate and intensify the common baseline worries experienced by the general population and place airmen at additional risks for mental health concerns and illness. This study assesses the effectiveness of the AF WebPHA mental health screening questions in predicting a mental health disorder diagnosis according to International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes generated by physicians or their surrogates. In order to assess the sensitivity, specificity, and positive predictive value of the AF WebPHA as a screening tool for mental health, survey results were compared to ascertain if they generated any mental health disorder related diagnosis for the period from January 1, 2009 to March 31, 2010. ^ Statistical analysis of the AF WebPHA mental health responses when compared with matching ICD-9-CM codes found that the sensitivity for 'Critical' or 'Priority' responses was only 3.4% and that it would correctly predict those who had the selected mental health diagnosis 9% of the time.^

Understanding treatment-related symptoms, symptom management techniques, and symptom management self-efficacy in a sample of non-muscle invasive bladder cancer patients

The current literature available on bladder cancer symptom management from the perspective of the patients themselves is limited. There is also limited psychosocial research specific to bladder cancer patients and no previous studies have developed and validated measures for bladder cancer patients’ symptom management self-efficacy. The purpose of this study was to investigate non-muscle invasive bladder cancer patients’ health related quality of life through two main study objectives: (1) to describe the treatment related symptoms, reported effectiveness of symptom-management techniques, and the advice a sample of non-muscle invasive bladder cancer patients would convey to physicians and future patients; and (2) to evaluate Lepore’s symptom management self-efficacy measure on a sample of non-muscle invasive bladder cancer patients. Methods. A total of twelve (n=12) non-muscle invasive bladder cancer patients participated in an in-depth interview and a sample of 46 (n=4) non-muscle invasive bladder cancer patients participated in the symptom-management self-efficacy survey. Results. A total of five symptom categories emerged for the participants’ 59 reported symptoms. Four symptom management categories emerged out of the 71 reported techniques. A total of 62% of the participants’ treatment related symptom-management techniques were reported as effective in managing their treatment-related symptoms. Five advice categories emerged out of the in-depth interviews: service delivery; medical advice; physician-patient communication; encouragement; and no advice. An exploratory factor analysis indicated a single-factor structure for the total population and a multiple factor structure for three subgroups: all males, married males, and all married participants. Conclusion. These findings can inform physicians and patients of effective symptom-management techniques thus improving patients’ health-related quality of life. The advice these patients’ impart can improve service-delivery and patient education.^

Cancer perceptions: Implications from the 2007 Health Information National Trends Survey

Background. Research has demonstrated associations between sociodemographic characteristics and illness perceptions; however, the impact of cancer exposure through personal or family diagnoses is not well-studied. The purposes of this study were to examine the prevalence of different cancer beliefs and the disparity in cancer beliefs across groups of individuals with distinct cancer histories; and to identify whether, when adjusted for sociodemographic characteristics, cancer history predicts a set of cancer beliefs.^ Methods. Using Leventhal’s Common Sense Model and data from the 2007 Health Information National Trends Survey (N=7172), we constructed multivariable logistic regressions to evaluate the effect of different stimuli, including cancer experience, on cancer perceptions (e.g., risk, worry, causation, outcome).^ Results. Findings indicate significant associations between cancer history and cancer perceptions. Individuals with family and personal cancer histories were more likely than individuals without any cancer history to worry about getting cancer (OR=3.55, P<0.01), agree they will develop cancer in the future (OR=8.81, P<0.01), and disagree that cancer is most often caused by a person’s behavior or lifestyle (OR=1.24, P=0.03). Additionally, results support education’s role in forming cancer perceptions. Individuals with high levels of education were more likely to endorse cancer prevention (OR=1.68, P<0.01) and higher 5-year survival rates (OR=1.41, P<0.01). ^ Conclusions. Results indicate cancer history affects cancer perceptions throughout the cancer continuum. Additionally, cancer history may influence coping behaviors and outcomes related to cancer.^ Impact. Cancer education and survivorship programs should assess important variables (e.g., cancer history) to more effectively tailor services and monitor evolving needs throughout cancer care.^

HPV vaccine coverage in Texas counties: An application of multilevel, small area estimation

Health departments, research institutions, policy-makers, and healthcare providers are often interested in knowing the health status of their clients/constituents. Without the resources, financially or administratively, to go out into the community and conduct health assessments directly, these entities frequently rely on data from population-based surveys to supply the information they need. Unfortunately, these surveys are ill-equipped for the job due to sample size and privacy concerns. Small area estimation (SAE) techniques have excellent potential in such circumstances, but have been underutilized in public health due to lack of awareness and confidence in applying its methods. The goal of this research is to make model-based SAE accessible to a broad readership using clear, example-based learning. Specifically, we applied the principles of multilevel, unit-level SAE to describe the geographic distribution of HPV vaccine coverage among females aged 11-26 in Texas.^ Multilevel (3 level: individual, county, public health region) random-intercept logit models of HPV vaccination (receipt of ≥ 1 dose Gardasil® ) were fit to data from the 2008 Behavioral Risk Factor Surveillance System (outcome and level 1 covariates) and a number of secondary sources (group-level covariates). Sampling weights were scaled (level 1) or constructed (levels 2 & 3), and incorporated at every level. Using the regression coefficients (and standard errors) from the final models, I simulated 10,000 datasets for each regression coefficient from the normal distribution and applied them to the logit model to estimate HPV vaccine coverage in each county and respective demographic subgroup. For simplicity, I only provide coverage estimates (and 95% confidence intervals) for counties.^ County-level coverage among females aged 11-17 varied from 6.8-29.0%. For females aged 18-26, coverage varied from 1.9%-23.8%. Aggregated to the state level, these values translate to indirect state estimates of 15.5% and 11.4%, respectively; both of which fall within the confidence intervals for the direct estimates of HPV vaccine coverage in Texas (Females 11-17: 17.7%, 95% CI: 13.6, 21.9; Females 18-26: 12.0%, 95% CI: 6.2, 17.7).^ Small area estimation has great potential for informing policy, program development and evaluation, and the provision of health services. Harnessing the flexibility of multilevel, unit-level SAE to estimate HPV vaccine coverage among females aged 11-26 in Texas counties, I have provided (1) practical guidance on how to conceptualize and conduct modelbased SAE, (2) a robust framework that can be applied to other health outcomes or geographic levels of aggregation, and (3) HPV vaccine coverage data that may inform the development of health education programs, the provision of health services, the planning of additional research studies, and the creation of local health policies.^

Analysis of participation in the federally mandated coal workers' health surveillance program

The Federal Coal Mine Health and Safety Act of 1969 required that periodic chest radiographs be offered to underground coal miners to protect the miners from the development of Coal Workers' Pneumoconiosis (CWP) and progression of the disease to progressive massive fibrosis (PMF). These examinations are administered by the National Institute for Occupational Safety and Health (NIOSH) through the Coal Workers' Health Surveillance Program (CWHSP). The mine operator is required to provide each miner with the opportunity to have the chest radiograph at no cost to the miner.^ Three rounds of examinations have been conducted since 1969 and the fourth is underway. The decrease in participation over rounds is of great concern if the incidence and progression of CWP are to be understood and controlled.^ This study developed rates of participation for each of 558 West Virginia underground coal mines who submitted or had NIOSH assigned plans for making chest radiographs available during the third round, July 1978 through December 1980. These rates were analyzed in relation to desired levels of participation and to reinforcing, predisposing and enabling factors presumed to affect rates of participation in disease prevention and surveillance programs.^ Two reinforcing factors, size of mine and inclusion of the mine in the National Coal Study (NCS) epidemiology research program, and the enabling factor, use of an on-site radiograph facility, demonstrated highly significant relationships to participation rates.^ The major findings of the study were: (1) Participation in the CWHSP is even lower than previously estimated; (2) CWHSP program evaluation is not systematic and program data base is not complete and comprehensive; and (3) NIOSH program policy is not clear and administration of the CWHSP is fragmented and lacks adequate fiscal and personnel resources. ^

Social network size and perceived network sufficiency as predictors of perceived health status

This study focused on the relationship between social network size (number of friends and relatives), perceived sufficiency of the network and self-rated health utilizing data from the National Survey of Personal Health Practices and Consequences, 1979. For men neither perceived sufficiency nor number of relatives were associated with self-rated health status. The number of friends was positively associated with health status. For women perceived network sufficiency was positively and significantly related to health status, independent of network size. The number of friends and relatives was not associated with self-rated health status. The sociodemographic variables accounted for most of the explained variance in health status for both males and females. Social networks may hold different meanings for women and men, and may require qualitative as well as quantitative analysis. There may have been insufficient variance in the major variables to produce meaningful results. ^

Community diagnosis: A working tool for developing culturally appropriate diarrhea intervention programs

Most studies related to diarrhea have been focused narrowly on the etiological and pathophysiological factors involved in inducing the disease. Such studies have often failed to consider other facets contributing to and possibly prolonging the problem, namely: socio-economic conditions, educational opportunities and attainments, cultural characteristics and beliefs, and the political administration and its commitment towards fulfilling its moral and ethical obligations in responding to and fostering human development.^ This study utilized the diagnostic approach of the PRECEDE model. The acronym stands for the predisposing reinforcing, and enabling constructs in educational diagnosis and evaluation. The constituents of this model were identified by utilizing participant observation field methods, traditionally applied by ethnographers to collect data describing the multiple facets of a culture, and linguistic anthropology used to capture and define characteristic semantics and viewpoints. The household study sample was randomly selected from a defined list of households known to have at least one child less than five years of age. An open-ended questionnaire format was used to interview the 115 mothers in the selected households sample.^ Kalama, the study community, is characteristically an agricultural village, situated in the Governorate of Kaliobia and located approximately 25 Km (about 15.5 miles) from the capital, Cairo, Egypt. The 1986 census indicates a population size of 13,328 people in 4,818 households. There were 65 deaths occurring among children less than five years in 1986. The causes of death were primarily related to diarrhea, followed by upper respiratory infections, congenital anomalies and birth injuries.^ This study outlines (a) practices related to the management of diarrhea, including the administration of foods and drinks during such episodes; (b) influences of governmental policies; and (c) recommended strategies for overcoming barriers and promoting effective diarrhea intervention programs. ^

Factors associated with home health care costs

Characteristics of Medicare-certified home health agencies in Texas and the contributions of selected agency characteristics on home health care costs were examined. Cost models were developed and estimated for both nursing and total visit costs using multiple regression procedures. The models included home health agency size, profit status, control, hospital-based affiliation, contract-cost ratio, service provision, competition, urban-rural input-price differences, and selected measures of patient case-mix. The study population comprised 314 home health agencies in Texas that had been certified at least one year on July, 1, 1986. Data for the analysis were obtained from Medicare Cost Reports for fiscal year ending between July 1, 1985 to June 30, 1986.^ Home health agency size, as measured by the logs of nursing and total visits, has a statistically significant negative linear relationship with nursing visit and total visit costs. Nursing and total visit costs decrease at a declining rate as size increases. The size-cost relationship is not altered when controlling for any other agency characteristic. The number of visits per patient per year, a measure of patient case-mix, is also negatively related to costs, suggesting that costs decline with care of chronic patients. Hospital-based affiliation and urban location are positively associated with costs. Together, the four characteristics explain 19 percent of the variance in nursing visit costs and 24 percent of the variance in total visit costs.^ Profit status and control, although correlated with other agency characteristics, exhibit no observable effect on costs. Although no relationship was found between costs and competition, contract cost ratio, or the provision on non-reimburseable services, no conclusions can be made due to problems with measurement of these variables. ^

A methodologic study of pain assessment in Anglo and Hispanic children with cancer

There is growing clinical evidence that even young children experience pain and accompanying anxiety. Few instruments have been validated to assess pain characteristics in children. The study of related demographic, illness, psychologic and parental factors in children has also been limited. This study examines the reliability and validity of pain assessment tools in an outpatient pediatric cancer population. A total of 78 children from three to fifteen years of age were observed and interviewed about the pain of invasive procedures. The effect of cultural factors and the stress of acculturation were examined by comparing data from two cultural groups, Anglo and Hispanic.^ Spielberger State-Trait Anxiety Scales were administered to children and parents prior to an invasive procedure. The Procedure Behavioral Checklist (PBCL) was used for observation of the child's response during the procedure. The Children's Procedural Interview (CPI) which contains items on the PBCL and visual analogues (scales of faces indicating varying degrees of pain and anxiety) was administered following the procedure.^ Reliability coefficients for Anglos were.78 on the PBCL,.79 on the CPI and.85 on the visual analogue scales. For Hispanics, the reliability for the PBCL was.54, while the CPI had a reliability of.72 and the visual analogue scales,.87. Construct validity was demonstrated by high correlations between the PBCL and CPI scores for both ethnic groups (.66 for Anglos and.64 for Hispanics) and by the significant correlation of State anxiety scores with both PBCL and CPI scores. Age was inversely correlated with PBCL and CPI scores for both ethnic groups. Hispanic parents' anxiety scores were higher than Anglo parents, but were not highly correlated with their child's PBCL, CPI or State-Trait anxiety scores. Caregivers' ratings were correlated with the PBCL scores for Anglos but not for Hispanics.^ The findings of this study indicate that pain responses may be reliably assessed using both observational and self-report methods in children, though differences in Anglo and Hispanic cultures exist. Differences in pain symptomatology and assessment in the two cultural groups warrant further study. ^

A study of the determinants of the utilization of preventive health care service among Chicanos

This study examines the individual and health care system determinants of two types of preventive health care practice behaviors, having a routine physical exam or a preventive dental exam, in the past year among Chicanos in the Southwestern United States. The study utilizes the Health System Model, developed by Aday and Andersen in 1974, to analyze the relative effect of education, income and occupation on the use of discretionary health care, controlling for other individual and health care system determinants.^ The study is based on a sample of 4,111 Mexican origin adults, drawn from the Hispanic Health and Nutrition Examination Survey (HHANES). This sample is representative of Mexican American residing in the Southwestern United States.^ The study tests the hypothesis that education is the most important social class predictor of preventive health care practice behavior. The fully elaborated model tests the hypothesis that individual determinants alone are insufficient to explain the use of preventive health care services among Chicanos.^ The study found that education and income are statistically significant social class indicators only as it relates to having a preventive dental exam. Education is not the most important social class predictor of either preventive health care practice behavior. Health care system determinants are key predictors of both behaviors. Need, as measured by self-perceived health status of teeth and gender, is as important a determinant as having dental insurance coverage as it relates to having a preventive dental exam. Implications for health programs to effectively reach Chicano target groups and remove access barriers to their use of discretionary health care services are discussed. ^

Patient satisfaction in government health facilities in the state of Qatar

This study provides data which can contribute to improving services and delivering quality health care in government health facilities in the state of Qatar. To measure the satisfaction with current care of selected patients who receive care in Hamad General Hospital and the Khalifa Town Health Center in the city of Doha, a cross-sectional survey and a self-administered questionnaire were used.^ Analysis was performed on data from 444 patients on eight dimensions of patient satisfaction with medical care. These include: general satisfaction, availability of services, convenience of services, facilities, humaneness of doctors, quality of care, continuity of care, and aspects of the last visit. Patient satisfaction parameters were compared for males vs. females, for citizens vs. non-citizens, and for patients seen in the hospital vs. those seen in the health center.^ Results indicate that patients seen in the hospital were more satisfied with care than patients seen in the health center, that non-citizens were more satisfied than citizens, and that males were slightly more satisfied than females with medical services. ^

Infectious diseases, immune status, and health insurance among child care providers in Houston, Texas

The objectives of this study were to compare female child-care providers with female university workers and with mothers of children in child-care centers for: (1) frequency of illness and work loss days due to infectious diseases, (2) prevalence of antibodies against measles, rubella, mumps, hepatitis B, hepatitis A, chickenpox and cytomegalovirus (CMV), and (3) status regarding health insurance and job benefits.^ Subjects from twenty child-care centers and twenty randomly selected departments of a university in Houston, Texas were studied in a cross-sectional fashion.^ A cluster sample of 281 female child-care providers from randomly selected child-care centers, a cluster sample of 286 university workers from randomly selected departments and a systematic sample of 198 mothers of children from randomly selected child-care centers.^ Main outcome measures were: (1) self-reported frequency of infectious diseases and number of work-days lost due to infectious diseases; (2) presence of antibodies in blood; and (3) self-reported health insurance and job benefits.^ In comparison to university workers, child-care providers reported a higher prevalence of infectious diseases in the past 30 days; lost three times more work-days due to infectious diseases; and were more likely to have anti-core antibodies against hepatitis B (odds ratio = 3.16 95% CI 1.27-7.85) and rubella (OR 1.88, 95% CI 1.02-3.45). Child-care providers had less health insurance and job-related benefits than mothers of children attending child-care centers.^ Regulations designed to reduce transmission of vaccine and non-vaccine preventable diseases in child-care centers should be strictly enforced. In addition policies to improve health insurance and job benefits of child-care providers are urgently needed. ^

Health status of Saudi women, the case of Yanbu al-Sinaiya

Detailed data on the health status of Saudi women are lacking. This cross sectional study attempts to provide a comprehensive description of the health status of Saudi females between the ages of 15-45 residing in Yanbu Al-Siniyah. The purpose is to assess women's needs for health services. The health status indicators are chronic tracer conditions, reported symptoms and multidimensional functioning levels. The generic functioning instrument of the Medical Outcome Study was used to estimate physical, social, and role functioning; degree of pain and health perceptions. The information was obtained by interviewing subjects and abstracting facts from their medical records. The results show functioning scores are in the "well health" range for physical, social, role and pain. Crowding and education have an equal or stronger effect of reducing functioning levels than the diagnosed tracer conditions. The highest prevalence conditions having a definite functional impact and diagnosed adequately in primary care are anemia, urinary tract infection, hemorrhoids, rheumatoid arthritis, caries and gingivitis. Reported symptoms strongly reducing function levels in this study are dyspnea, heart pain, incontinence, eye and skin problems, and joint ache. The impact of the reliability and validity of the measures used and other limitations of the results are discussed. Finally, some policy implications and suggestions for future study are presented. ^