Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Insurance Selection Bias”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

990 records · Page 55Linked to original sources
PubMed2001

Bias, discrimination, and obesity.

This article reviews information on discriminatory attitudes and behaviors against obese individuals, integrates this to show whether systematic discrimination occurs and why, and discusses needed work in the field. Clear and consistent stigmatization, and in some cases discrimination, can be documented in three important areas of living: employment, education, and health care. Among the findings are that 28% of teachers in one study said that becoming obese is the worst thing that can happen to a person; 24% of nurses said that they are "repulsed" by obese persons; and, controlling for income and grades, parents provide less college support for their overweight than for their thin children. There are also suggestions but not yet documentation of discrimination occurring in adoption proceedings, jury selection, housing, and other areas. Given the vast numbers of people potentially affected, it is important to consider the research-related, educational, and social policy implications of these findings.

Attitude of Health Personnel↗
PubMed2001

Factors used to select adjuvant therapy of breast cancer in the United States: an overview of age, race, and socioeconomic status.

Age, race, and socioeconomic status all play a role in decisions regarding breast cancer adjuvant therapy. Increasing age remains the major risk factor for breast cancer, while in very young women breast cancer may have a poorer prognosis, even when adjusted for disease stage and other variables. More than half of all new breast cancers in the United States occur in women older than 65 years. Because of the higher frequency of coexisting (comorbid) serious illness in older women, the benefits of adjuvant therapy get smaller as age increases. Adjuvant therapy with tamoxifen and/or chemotherapy can statistically significantly improve survival in older women, but older women are less likely to receive chemotherapy and are less likely to be offered participation in clinical trials. Efforts are now under way to overcome age bias among health care providers and to develop clinical trials focusing on older patients. Breast cancer mortality is higher in African-Americans than in white Americans. Although the biologic characteristics of breast cancer are worse in African-Americans, major differences in survival are related to socioeconomic factors and access to care. When matched for disease stage and other major clinical and biologic variables, African-American and white patients have similar survival rates. Few data are available on the effects of adjuvant treatment on early breast cancer outcomes in Hispanic Americans and Asian-Americans. Poverty and lack of insurance are surrogates for poor outcomes; major efforts are needed to guarantee all Americans high-quality cancer care.

Adult↗
PubMed1985

Biased selection in Twin Cities health plans.

The data in Tables 1 through 4 show significant differences in the enrollment of higher health-related financial risk individuals and their families among health plans. FFS enrollees are older and exhibit more chronic illness on average. IPAs enroll a greater proportion of females than do PGP or FFS plans. PGPs and IPAs do not differ significantly in the age and chronic illness of their enrollees, but IPAs enroll a significantly greater proportion of females than do PGPs. The age difference between FFS and prepaid plans appears to be greater for long-term enrollees. The same pattern is true of chronic illness, but the results are often not statistically significant. We do not have time-series data, however, and cannot conclude that future comparisons among long-term enrollees will remains as they are now. In any care our data do not support the hypothesis that biased selection is a short-term problem that will be corrected as the population in prepaid plans ages. Our data contain a cross-section of environments for health plans in firms: long- and short-term offerings, long- and short-term enrollees, high and low out-of-pocket premium costs, etc. Our strongest results are the simplest: across all plans and environments there are significant differences in enrollee characteristics. These differences would not be inefficient if all groups paid actuarially fair premiums. However, mandatory offering and community-rating allow prepaid plans to enroll a younger population with less chronic illness and to maintain an information asymmetry that prevents employers and employees from determining--either prior to or following enrollment--the relationship of the prepaid plan's premium to its marginal cost.

Age Factors↗
PubMed1981

Radiation detection and diagnosis of breast cancer.

The value of mammography in the symptomatic patient has been adequately documented, but its use as a detection procedure remains a question. Risk-benefit ratios, based primarily upon the study carried out by the Health Insurance Plan of Greater New York, have suggested that the technique has little value in individuals under age 50. Emphasis has been placed upon the possible carcinogenic effects of radiation as compared with the efficacy of mammography and the questionable influence of early diagnosis upon end results. Although technical advances have substantially reduced the exposure of the patient to radiation, the possibility of significant information loss as the result of these developments has been considered a potential drawback to their routine use. All of these factors have served to diminish both public and professional acceptance of the examination. Although current data do not allow complete resolution of these problems, certain conclusions may be drawn and trends established. The sum of these may indicate that minimal dose mammography is an accurate, low-risk procedure, capable of significantly altering the natural history of breast cancer. Whether or not the examination should be routinely used in women under age 50 remains open to question since the lack of experimental controls prohibits validation of the technique in terms of reduced mortality rates. Documentation of increased survival rates may partially assist in the established of a reliable risk-benefit ratio, but will not satisfy the statistical requirements of eliminating lead-bias, and self-selection. These questions may be resolved by studies now underway.

Adult↗
PubMed2003

Management of previous cesarean section.

PURPOSE OF REVIEW: The management of cesarean sections causes much controversy among healthcare providers, patients, and insurers. A trial of vaginal birth after previous cesarean is reported to be a safe and practical method to reduce the rate of cesarean sections. The popularity of vaginal birth after previous cesarean has increased over the past two decades, but rates have recently started to decline again. This review will evaluate recent literature that might be responsible for this reversal in trend. RECENT FINDINGS: Earlier studies on previous cesarean section pregnancies focused primarily on the success rate of vaginal birth after previous cesarean, which is reported to be 60-80%. Recent large, retrospective, population-based cohort studies examined the maternal and neonatal safety of trial of labour compared with elective repeat cesarean delivery, and confirmed that the risks of uterine rupture and neonatal mortality were significantly increased after trial of labour, particularly when induced with prostaglandins. However, the absolute risk of adverse events remains small. The maternal and neonatal morbidity risk increases when vaginal birth after previous cesarean attempts fails, which emphasizes the importance of careful case selection. SUMMARY: Recent studies highlighted the risks of attempted vaginal birth after previous cesareans, especially when trials fail, but have not addressed the long-term risks of an elective repeat cesarean delivery. The assessment of treatment risks by observational studies is subject to bias, because the different treatment groups may not be comparable at the outset. In the absence of better data, the counselling of such women must currently be based on this evidence.

Cesarean Section, Repeat↗
PubMed2005

Wheeze, asthma diagnosis and medication use: a national adult survey in a developing country.

BACKGROUND: As relatively little is known about adult wheeze and asthma in developing countries, this study aimed to determine the predictors of wheeze, asthma diagnosis, and current treatment in a national survey of South African adults. METHODS: A stratified national probability sample of households was drawn and all adults (>14 years) in the selected households were interviewed. Outcomes of interest were recent wheeze, asthma diagnosis, and current use of asthma medication. Predictors of interest were sex, age, household asset index, education, racial group, urban residence, medical insurance, domestic exposure to smoky fuels, occupational exposure, smoking, body mass index, and past tuberculosis. RESULTS: A total of 5671 men and 8155 women were studied. Although recent wheeze was reported by 14.4% of men and 17.6% of women and asthma diagnosis by 3.7% of men and 3.8% of women, women were less likely than men to be on current treatment (OR 0.6; 95% confidence interval (CI) 0.5 to 0.8). A history of tuberculosis was an independent predictor of both recent wheeze (OR 3.4; 95% CI 2.5 to 4.7) and asthma diagnosis (OR 2.2; 95% CI 1.5 to 3.2), as was occupational exposure (wheeze: OR 1.8; 95% CI 1.5 to 2.0; asthma diagnosis: OR 1.9; 95% CI 1.4 to 2.4). Smoking was associated with wheeze but not asthma diagnosis. Obesity showed an association with wheeze only in younger women. Both wheeze and asthma diagnosis were more prevalent in those with less education but had no association with the asset index. Independently, having medical insurance was associated with a higher prevalence of diagnosis. CONCLUSIONS: Some of the findings may be to due to reporting bias and heterogeneity of the categories wheeze and asthma diagnosis, which may overlap with post tuberculous airways obstruction and chronic obstructive pulmonary disease due to smoking and occupational exposures. The results underline the importance of controlling tuberculosis and occupational exposures as well as smoking in reducing chronic respiratory morbidity. Validation of the asthma questionnaire in this setting and research into the pathophysiology of post tuberculous airways obstruction are also needed.

Adolescent↗
PubMed2002

Nuptial gifts and the evolution of male body size.

In many insect systems, males donate nuptial gifts to insure an effective copulation or as a form of paternal investment. However, if gift magnitude is both body size-limited and positively related to fitness, then the opportunity exists for the gift to promote the evolution of large male size. In the striped ground cricket, Allonemobius socius, males transfer a body size-limited, somatic nuptial gift that is comprised primarily of hemolymph. To address the implications of this gift on male size evolution, we quantified the intensity and direction of natural (fecundity) and sexual (mating success) selection over multiple generations. We found that male size was under strong positive sexual selection throughout the breeding season. This pattern of selection was similar in successive generations spanning multiple years. Male size was also under strong natural selection, with the largest males siring the most offspring. However, multivariate selection gradients indicated that gift size, and not male size, was the best predictor of female fecundity. In other words, direct fecundity selection for larger gifts placed indirect positive selection on male body size, supporting the hypothesis that nuptial gifts can influence the evolution of male body size in this system. Although female size was also under strong selection due to a size related fecundity advantage, it did not exceed selection on male size. The implications of these results with regard to the maintenance of the female-biased size dimorphic system are discussed.

Analysis of Variance↗
PubMed1996

Motor vehicle crashes in roadway construction workzones: an analysis using narrative text from insurance claims.

Motor vehicle travel through roadway construction workzones has been shown to increase the risk of a crash. The number of workzones has increased due to recent congressional funding in 1991 for expanded roadway maintenance and repair. In this paper, we describe the characteristics and costs of motor vehicle crashes in roadway construction workzones. As opposed to using standard accident codes to identify accident types, automobile insurance claims files from 1990-93 were searched to identify records with the keyword "construction" in the accident narrative field. A total of 3,686 claims were used for the analysis of crashes. Keywords from the accident narrative field were used to identify five pre-crash vehicle activities and five crash types. We evaluated misclassification error by reading 560 randomly selected claims and found it to be only 5%. For each of four years, 1990-93, there was a total of 648,996,977 and 1,065 crashes, respectively. There was a 70% increase in the crash rate per 10,000 personal insured vehicles from 1990-93 (2.1-3.6). Most crashes (26%) involved a stopped or slowing vehicle in the workzone. The most common crash (31%) was a rear-end collision. The most costly pre-crash activity was a major judgment error on the part of a driver (n = 120, median cost = $2,628). An overturned vehicle was the most costly crash type (n = 16, median cost = $4,745). In summary, keyword text analysis of accident narrative data used in this study demonstrated its utility and potential for enhancing injury epidemiology. The results suggest interventions are needed to respond to growing traffic hazards in construction workzones.

Accidents, Traffic↗
PubMed2002

Equity in waiting times for major joint arthroplasty.

OBJECTIVE: To ascertain whether waiting lists are managed in an equitable fashion in a universal health system by examining demographic, socioeconomic and clinical factors, along with 2 health systems variables. DESIGN: A prospective survey by questionnaire. SETTING: The Capital Health Region of Edmonton, Alta. PATIENTS AND METHODS: A cohort of 553 patients, who were waiting for either total hip or total knee replacement surgery, seen between Dec. 18, 1995, and Jan. 24, 1997. INTERVENTIONS: A home visit was made when the patient was first placed on the waiting list and again just before surgery to complete the questionnaires. The Western Ontario and McMaster Universities (WOMAC) instrument and the Medication Quantification Score were administered at the time the patient was placed on the waiting list. MAIN OUTCOME MEASURE: The length of waiting time, defined as the date the patient was put on the waiting list to the date the patient was operated on. RESULTS: There were no biases in waiting time with respect to age, gender, education or work status. Although pain and function were not related to waiting time, multivariate analyses found that marital status, primary language, body mass index, pain medication use and the size of the surgeons' major joint replacement practice determined waiting time for surgery. However, this model explained only 10% of the variance in waiting time. CONCLUSION: Waiting lists were managed unfairly in terms of clinical equity (clinical severity) but managed fairly in terms of social equity.

Aged↗
PubMed1996

Estimating nonparticipation bias in a longitudinal study of bereavement.

Nonparticipants in epidemiological studies may differ in important respects from participants but the magnitude of this potential bias is rarely quantified. This study estimates the effect of nonparticipation on estimates of mental health problems following stillbirth, neonatal death or sudden infant death syndrome. Of 805 families approached, 512 (64 per cent) were recruited, of whom 77 per cent of mothers and 71 per cent of fathers completed four study interviews. Younger, unmarried, unemployed parents without private health insurance were less often recruited, and even if recruited, were less likely to complete the interview. By evaluating several possible scenarios, we estimated that had mothers lost to follow-up remained in the study, anxiety rates would have varied by no more than +/-4 per cent. Relative risks associated with bereaved-control comparisons would have differed little from the observed estimate of 2.33. Estimating the effects of initial nonresponse is more difficult but the adoption of a worst-case scenario produced a relative risk of 3.47. Despite systematic nonparticipation suggestive of social disadvantage, attrition-related bias may have had only a modest effect on anxiety and depression rate estimates. However, this may not be the case when sample loss is high, when associations between attrition and outcome are strong, and when attrition-related behaviour differs across comparison groups.

Adult↗
PubMedDate not supplied

The reliability and validity of birth certificates.

OBJECTIVES: To summarize the reliability and validity of birth certificate variables and encourage nurses to spearhead data improvement. DATA SOURCES: A Medline key word search of reliability and validity of birth certificate, and a reference review of more than 60 articles were done. STUDY SELECTION: Twenty-four primary research studies of U.S. birth certificates that involved validity or reliability assessment. DATA EXTRACTION: Studies were reviewed, critiqued, and organized as either a reliability or a validity study and then grouped by birth certificate variable. DATA SYNTHESIS: The reliability and validity of birth certificate data vary considerably by item. Insurance, birthweight, Apgar score, and delivery method are more reliable than prenatal visits, care, and maternal complications. Tobacco and alcohol use, obstetric procedures, and delivery events are unreliable. Birth certificates are not valid sources of information on tobacco and alcohol use, prenatal care, maternal risk, pregnancy complications, labor, and delivery. CONCLUSIONS: Birth certificates are a key data source for identifying causes of increasing U.S. infant mortality but have serious reliability and validity problems. Nurses are with mothers and infants at birth, so they are in a unique position to improve data quality and spread the word about the importance of reliable and valid data. Recommendations to improve data are presented.

Alcohol Drinking↗
PubMed2002

Subgroups of refusers in a disability prevention trial in older adults: baseline and follow-up analysis.

OBJECTIVES: This study explored differences between refusers and participants in a longitudinal study with extensive baseline and follow-up information. METHODS: Results of a trial comparing 791 participants and 401 community-residing older adults who refused to participate in a study concerning preventive home visits were examined. Information was collected from interviews, insurance records, and government files. RESULTS: Despite similarities in terms of age, sex, and self-perceived health at baseline, 3-year follow-up data indicated that refusers had a 1.58-fold higher risk of entering a nursing home than participants. There were additional differences between refusers and participants when refusers were categorized in 4 subgroups based on self-reported reason for refusal (too ill, too healthy, no interest, and other reasons). CONCLUSIONS: Future studies should include follow-up data to allow comparisons between refusers and participants and should address the presence of multiple subgroups of refusers.

Activities of Daily Living↗
PubMed1992

Vocational rehabilitation of cancer patients.

Indeed, cancer patients do face job discrimination. Empirical studies have shown evidences of personnel selection bias and many other work-related problems. Laws and legislation are good tools but are not the answers to most of the problems. A competent multidisciplinary rehabilitation team, a well-educated general public and employers, together with empowered cancer patients can do more towards eliminating destructive road blocks to the employment of the employable. This takes time and initially may be costly for all concerned. However, evidence has shown that the cost of closed cancer rehabilitation cases is the lowest per case for all occupational disabilities including tuberculosis, diabetes, mental disorders, heart disease, and orthopedic problems. In short, we dare not forget the simple truth put forth by Harold Dodds that "No, work is not an ethical duty imposed on us from without by a misguided and outmoded Puritan morality; it is a manifestation of man's deepest desire that the days of his life shall have significance." Cancer patients living with cancer desire significance to their days to the same degree that those of us who are healthy and well desire and need it.

Employment↗
PubMed2004

Recruitment into a long-term pediatric asthma study during emergency department visits.

OBJECTIVE: Asthma is the most common chronic illness in childhood. Recruiting children and their parents into a research study in a busy urban emergency department (ED) is challenging. The purpose of this manuscript is to describe the recruitment process and the results of our recruitment in soliciting children and their parents to participate in an ED-based asthma research study. METHODS: The data for this manuscript came from a National Heart, Lung, and Blood Institute-funded study: Study of Asthma Follow-up from the Emergency Department (SAFE). SAFE evaluated an ED-based intervention to link low-income urban children with asthma to their primary care providers. Two persons were assigned specifically to enrolling, which was done from 0700 to 2300 hours Monday through Friday. Data for the analysis come from the web-based database, the master log, and the hospital's patient database. A computerized randomization scheme chose 512 patients from all patients in the master log for more detailed demographic analyses. RESULTS: Five hundred twenty-seven subjects were enrolled between February 1999 and May 2001. There were 9188 children who presented for treatment of an acute asthma exacerbation during this interval. The number of eligible parents was similar to the predicted number. Chart reviews were conducted on a subset of patients presenting to the ED to ensure that the recruitment strategy did not bias the patients enrolled. Demographic characteristics of asthma patients were similar during enrollment and non-enrollment times. Comparison of patients who were enrolled with those who were not enrolled indicated no differences by gender, race, insurance status, age, or socioeconomic status of neighborhood residence. DISCUSSION: The high rate of enrollment was primarily due to the two dedicated enrollers. The enrollers quickly learned how to function within the ED and how to interact with both families and ED staff. Strategies identified by the enrollers as helpful in randomizing subjects included visits with the parents shortly after the physician had initiated treatment so that stability of the child had been achieved. Interacting with the child and showing concern for the comfort of both the child and parent during the ED stay were important as well. CONCLUSION: Recruiting subjects into long-term follow-up studies in the ED setting is a departure from traditional ED studies. The ED enrollment offers the distinct advantage of capturing subjects who are unlikely to present for care in other locations. We were able to successfully recruit low-income urban parents of children with asthma for study of both short-term and long-term outcomes. Careful attention to planning and then integration of enrollers into the ED setting can result in successful recruitment of patients and their parents. The enrollment process successfully captured the subjects of interest without bias.

Asthma↗
PubMed1997

Correlates and effect of non-response in a postpartum survey of obstetrical care quality.

This study investigated the differences between unprompted respondents, prompted respondents, and non-respondents to a postpartum postal survey, and determined the likely impact of non-response on the accuracy of calculations of patient assessments of obstetrical care quality. Birth certificate and hospital discharge data were obtained for 1664 live births at three hospitals in Washington State between 8/91-10/91 and linked with 1268 completed postpartum maternal postal surveys. Non-white race, public insurance payer, unmarried status, and smoking in pregnancy were independent risk factors for non-participation. Among participants, non-white race, unmarried status, and having an infant who was low birthweight, preterm, or discharged late were independent risk factors for prompted response. The inclusion of prompted respondents did not substantially alter the calculated proportion of women rating obstetrical care quality as low, and these figures were similar to proportions estimated for the entire intended cohort using a modification of Drane's method. A one-time mailing of an obstetrical care quality survey can provide information similar to that obtained with more extensive follow-up even though substantial differences may exist between unprompted and prompted respondents, and with adjustment for factors related to non-participation and timing of response, it may be possible to obtain accurate estimation of outcome prevalences for the entire intended cohort.

Adult↗
PubMed2001

General Practitioners' views on the provision of nicotine replacement therapy and bupropion.

BACKGROUND: Nicotine replacement therapies (NRT) and a new drug, bupropion, are licensed in several countries as aids to smoking cessation. General practitioners (GPs) play a crucial role in recommending or prescribing these medications. In the UK there has been discussion about whether the medications should be reimbursable by the National Health Service (NHS). This study assessed English GPs' attitudes towards reimbursement of NRT and bupropion. METHODS: Postal survey of a randomly selected national sample of GPs; 376 GPs completed the questionnaire after one reminder; effective response rate: 53%. There was no difference between the responses of GPs who responded to the initial request and those who responded only after a reminder suggesting minimal bias due to non-response. RESULTS: Attitudes of GPs were remarkably divided on most issues relating to the medications. Forty-three percent thought that bupropion should not be on NHS prescription while 42% thought that it should be (15% did not know); Fifty percent thought that NRT should not be on NHS prescription while 42% thought it should be (8% did not know). Requiring that smokers attend behavioural support programmes to be eligible to receive the medications on NHS prescription made no appreciable difference to the GPs' views. GPs were similarly divided on whether having the medications reimbursable would add unacceptably to their workload or offer a welcome opportunity to discuss smoking with their patients. A principal components analysis of responses to the individual questions on NRT and bupropion revealed that GPs' attitudes could be understood in terms of a single 'pro-con' dimension accounting for 53% of the total variance which made no distinction between the two medications. CONCLUSIONS: GPs in England appear to be divided in their attitudes to medications to aid smoking cessation and appear not to discriminate in their views between different types of medication or different aspects of their use. This suggests that their attitudes are generated by quite fundamental values. Addressing these values may be important in encouraging GPs to adhere more closely to national and international guidelines.

Antidepressive Agents, Second-Generation↗
PubMedDate not supplied

Is BMD testing appropriate for all menopausal women?

The United States Preventive Services Task Force has provided an evidence-based guideline indicating that bone mineral density (BMD) testing is appropriate for all women aged 65 or older. This does not preclude BMD testing in younger postmenopausal women but places the onus on the treating physician to justify the procedure to the patient and often the patient's insurance carrier. There are very few circumstances in which BMD testing is appropriate for healthy premenopausal women, but BMD testing in younger postmenopausal women is often appropriate: when there is a family history of osteoporosis with fracture, a personal history of fracture as an adult, and a medical, surgical or therapeutic history that might be associated with accelerated bone loss or increased risk of fracture. Medical conditions include intestinal diseases associated with malabsorption, such as non-tropical sprue, or primary hyperparathyroidism. Women who have neurologic conditions that increase the risk of falling should also be tested. There are data to suggest that patients with hemoglobinopathy are at increased risk for osteoporosis. Surgical conditions include the increasingly performed surgery for obesity and other surgery resulting in bowel resection (e.g., for inflammatory bowel disease). The major medication-related concern is corticosteroid therapy, but chronic or over-treatment with thyroxine, and chronic heparin therapy, should also be considered risk factors for osteoporosis. When performing a BMD test for the first time, it is essential to remember that 50% of women at menopause will have a negative T-score, but this does not imply that the patient has indeed lost any bone from her peak bone mass.

Absorptiometry, Photon↗
PubMed2006

Utilization and costs for compliant patients initiating therapy with pioglitazone or rosiglitazone versus insulin in a Medicaid fee-for-service population.

OBJECTIVE: To compare health care utilization and costs for type 2 diabetes patients initiating therapy with one of two thiazolidinediones (TZDs, pioglitazone or rosiglitazone) or insulin in a Medicaid population. METHODS: The study used a retrospective cohort design and included type 2 diabetes patients who initiated therapy with TZDs or insulin treatment during the 3-year period (1999-2001). These patients were identified from a Medicaid administrative claims database for approximately 230,000 fee-for-service Medicaid recipients using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes: 250.0x-250.9x, where x = 0 or 2). The first pharmacy claim for a TZD or insulin was treated as an index pharmacy claim, and utilization and costs were assessed for each patient for a 12-month follow-up period after the index date of the first pharmacy claim. A 12-month preperiod without a pharmacy claim for a TZD or insulin confirmed that the patient was newly prescribed with these medications. Analysis was restricted to a compliant sample receiving at least 6 pharmacy claims for either TZD or insulin in the 12-month follow-up period. The propensity matching technique was used to control for selection bias and potential imbalances between these groups at baseline. Patients initiating therapy with insulin or TZD were matched on the basis of demographics, year of index pharmacy claim, presence of microvascular/macrovascular complications, comorbidity, type 2 diabetes-related medical utilization and costs in the 12-month preperiod, overall health care utilization and costs in the preperiod, and type of oral hypoglycemic agents/other medications in the preperiod. Nonparametric bootstrapping was used to estimate the impact of therapy on both overall and type 2 diabetes-related health care utilization and costs in the matched sample. RESULTS: A total of 2,842 patients with type 2 diabetes patients met the inclusion criteria prior to exclusion of 881 patients (31.0%) who did not receive at least 6 pharmacy claims for either TZD or insulin in the 12 months following the index pharmacy claim, leaving 1,961 type 2 diabetes patients who initiated therapy with one of the two TZDs or insulin in the 3-year enrollment period (TZDs = 1,523; insulin = 438). Propensity matching eliminated 1,271 patients (64.8%), resulting in a final sample consisting of 690 patients.345 patients per treatment group.with comparable demographic and utilization parameters at baseline. In the 12-month follow-up period for the measures of overall utilization, patients initiated on TZDs did not differ significantly in the number of emergency room (ER)/hospitalization episodes and the number of pharmacy claims compared with the patients initiated on insulin, but they did have an average of 1.2 fewer physician office visits (9.3 vs. 10.5, P <0.05). Compared with the insulin group, the TZD group incurred 35% lower costs for ER visits/hospitalization (dollar 3,727 vs. dollar 5,793, P <0.01) and 18% lower total health care costs (dollar 12,737 vs. dollar 15,563, P <0.05). No significant differences were observed between the groups in overall outpatient and pharmacy costs. When the analysis was restricted to type 2 diabetes-related utilization and costs, patients initiating therapy with TZDs had 0.8 fewer physician office visits (3.6 vs. 4.4, P <0.05). However, TZD patients had 2.7 (14.1%) more diabetes pharmacy claims than patients initiating therapy with insulin (P <0.01), but there was no difference in the number of type 2 diabetes-related ER visits/hospitalizations between the groups. The TZD group had 53% higher type 2 diabetes-related pharmacy costs than the insulin group (dollar 1,678 vs. dollar 1,096, P <0.01). However, these costs were offset by lower costs for ER visits and hospitalization for the TZD group as compared with the insulin group (dollar 2,855 vs. dollar 5.090, P <0.01) resulting in 25% lower total type 2 diabetes-related costs for the TZD group compared with the insulin group (dollar 5,425 vs. dollar 7,255, P <0.05). CONCLUSION: Medicaid fee-for-service patients initiated on either pioglitazone or rosiglitazone incurred higher diabetes-related pharmacy costs, which were offset by lower total type 2 diabetes-related medical costs, contributed primarily by lower costs for ER visits and hospitalizations in this 12-month analysis.

Diabetes Mellitus, Type 2↗