Risk Adjustment, Mode Adjustment, and Nonresponse Bias Analysis on Quality Measures From a Long-Term Care Hospital Experience of Care Survey.

OBJECTIVES: To develop a patient risk adjustment model for experience of care (EOC) quality measures for long-term care hospitals (LTCHs) that includes mode of survey administration. To assess presence of nonresponse bias in the adjusted facility-level scores. DESIGN: We tested 3 modes of collecting the EOC data: mail-only, mixed (ie, mail with telephone follow-up), and in-facility. This study used sequential modeling and impact analysis, specified a risk and mode adjustment model, and evaluated presence of nonresponse after adjustment. SETTING: LTCHs. PARTICIPANTS: Patients (N=1364) and 69 LTCHs. INTERVENTION: Not applicable. MAIN OUTCOME MEASURES: Risk and mode adjusted responses to 28 survey questions and 6 facility-level scores derived from survey responses. RESULTS: Mode of data collection and patient risk variables (age, sex, overall health, overall mental health, marital status, education, race, and whether a proxy responded) were included in the model. Clinical variables were not significant. The in-facility mode was associated with significantly higher performance scores than the other modes. When the recommended risk and mode adjustment model was applied, nonresponse bias was not observed in any mode. CONCLUSIONS: LTCH EOC data should be adjusted for patient risk variables including mode of data collection.

Findings From Implementing a Patient Experience Survey in a Quality Measurement System for Substance Abuse Disorder Treatment Facilities in 6 States.

OBJECTIVES: Patient experience surveys (PESs) are an important component of determining the quality of health care. There is an absence of PES data available to people seeking to identify higher quality substance use disorder treatment providers. Our project aimed to correct this by implementing a PES for substance use disorder treatment providers and publicly disseminating PES information. METHODS: We created a population frame of all addiction providers in 6 states. Providers were asked to disseminate a survey invitation letter directing patients to a survey Web site. No personally identifiable information was exchanged. We developed a 10-question survey, reflecting characteristics National Institute on Drug Abuse (NIDA), National Institute on Alcohol Abuse and Alcoholism (NIAAA), Substance Abuse and Mental Health Services Administration (SAMHSA) have identified as reflecting higher-quality addiction treatment. RESULTS: Nineteen percent of facilities participated; among participating facilities, 9627 patients completed the survey. Patient experience varied significantly by facility with the percentage of a facility's patients who chose the most positive answer varying widely. We calculated that between-facility reliability will meet or exceed 0.80 for facilities with 20 or more responding patients. We searched for but did not find evidence of data falsification. CONCLUSIONS: This cost-efficient survey protocol is low burden for providers and patients. Results suggest significant differences in quality of care among facilities, and facility-level results are important to provide to consumers when they evaluate the relative patient-reported quality of facilities. The data are not designed to provide population-based statistics. As more facilities and patients per facility participate, public-facing PES data will be increasingly useful to consumers seeking to compare and choose facilities.

Comparison of patient visits to emergency departments, physician offices, and dental offices for dental problems and injuries.

OBJECTIVES: Our understanding of the use of emergency departments (EDs) and physician offices for the management of dental problems is limited. We undertook this study to examine whether there are differences in their use by low-income White and minority adults as compared with higher-income adults. METHODS: Participantsincluded White, Black, and Hispanic adults who had experi enced a dental problem during the previous 12 months and who visited a physician, ED, or dentist for treatment. We selected a stratified random sample of 27,002 Maryland households with listed telephones to screen for eligibility. We identified 1,387 households with an eligible adult, selected 423 for interviews, and completed interviews with 401 (94.8%). RESULTS: To restore correct proportionality to the sample, and to adjust for nonresponse and the distribution of demographic characteristics, weights were created for use in the analyses. Only 7.1 percent of respondents contacted an ED, while 14.3 percent contacted a physician and 90.2 percent a dentist. The vast majority of respondents who contacted an ED (96.0%) or a physician (92.2%) also contacted a dentist. Lower-income respondents were more likely to seek care from an ED, while higher-income respondents were more likely to seek care from a dentist. Over whelmingly, respondents visiting EDs (89.4%) and physicians (51.7%) were instructed to see a dentist or given prescriptions/samples. Treatment provided by EDs, physicians, and dentists was not associated with the respondent's income or race/ethnicity. CONCLUSIONS: Respondents visiting EDs and physicians typically did not receive definitive care and subsequently visited a dentist for treatment.

Comparison of patient centeredness of visits to emergency departments, physicians, and dentists for dental problems and injuries.

BACKGROUND: Patient-centered care has a positive impact on patient health status. This report compares patient assessments of patient centeredness during treatment in hospital emergency departments (EDs) and physician and dentist offices for dental problems and injuries. RESEARCH DESIGN: Participants included low-income White, Black, and Hispanic adults who had experienced a dental problem or injury during the previous 12 months and who visited an emergency department, physician, or dentist for treatment. A stratified random sample of Maryland households participated in a cross-sectional telephone survey. Interviews were completed with 94.8% (401/423) of eligible individuals. Multivariable logistic regression analyses were performed. RESULTS: The measure of predictive power, the pseudo-R2s, calculated for the logistic regression models ranged from 12% to 18% for the analyses of responses to the measures of patient centeredness (satisfaction with treatment, careful listening, thorough explaining, spending enough time, and treated with courtesy and respect). EDs were less likely than dentists to treat patients with great courtesy and respect. CONCLUSIONS: Further research is needed to identify factors that support patient-centered care.

More accurate racial and ethnic codes for Medicare administrative data.

Analyses of health care disparities in Medicare using administrative race and ethnicity data have typically been limited to Black and White beneficiaries. This is in part due to the small size of the other categories, inaccuracies in the race and ethnicity codes, and caveats that more extensive analyses would produce biased results. While previous Medicare efforts certainly improved the accuracy of race and ethnicity coding, we have developed an imputation algorithm that dramatically improves the accuracy of coding for Hispanic and Asian or Pacific Islander beneficiaries. When compared with self-reported race and ethnicity, sensitivity increased from 29.5 to 76.6 percent for Hispanic and from 54.7 to 79.2 percent for Asian and Pacific Islander beneficiaries, with no loss of specificity, and Kappa coefficients reaching 0.80. As a result, 2,245,792 beneficiaries were recoded to Hispanic and 336,363 to Asian or Pacific Islander.

Impact of influenza immunization on medical expenditures among Medicare elderly, 1999-2003.

BACKGROUND: The impact of influenza immunization on expenditures for inpatient, outpatient, and professional services among elderly Medicare beneficiaries between 1999 and 2003 was examined. METHODS: Data were from independent annual survey samples of 175,000 beneficiaries. Response rates ranged from 64% to 71%. Survey data included beneficiaries' demographics, education, supplemental insurance, perceived health, and influenza vaccination. Baseline measures, derived from Medicare claims for the year prior to influenza season, included service utilization, comorbidities, influenza immunization, and health status. The outcome measure was medical expenditures for acute and chronic respiratory conditions (ACRCs) for each 33-week annual influenza season. RESULTS: Total expenditures for ACRCs were lower among the immunized population during all four seasons. The amount and statistical significance of the savings varied with the severity of the virus and the vaccine match to the prevalent influenza strains. During the 1999-2000 influenza season, which had the most severe virus and a close vaccine match, average costs for ACRCs were $88 lower among immunized beneficiaries than among nonimmunized beneficiaries (equivalent to a 3.06% savings). During the 2002-2003 season, which had a less severe virus but the highest vaccine match rate, average costs for ACRCs were $103 lower for immunized beneficiaries than for nonvaccinated beneficiaries (equivalent to a 3.12% savings). The relative reduction in ACRC expenditures among vaccinated beneficiaries is attributable to less frequent use of inpatient services. CONCLUSIONS: In addition to improving the health of older Americans, meeting the Healthy People 2010 influenza immunization goal of 90% among the elderly should also result in lower Medicare expenditures.

Disparities in immunizations among elderly Medicare beneficiaries, 2000 to 2002.

BACKGROUND: Research reveals that influenza and pneumococcal immunization rates among blacks, Asians, and Hispanics significantly trail those of whites. This 2003 study examines recent trends and disparities for influenza and pneumococcal immunizations among elderly, non-institutionalized Medicare beneficiaries. METHODS: National samples of approximately 179,000 Medicare fee-for-service beneficiaries were surveyed by mail and telephone each year from 2000 to 2002. Outcomes include self-reported influenza immunization in the previous year and receipt of a pneumococcal immunization ever. RESULTS: Influenza immunization dipped in 2001 (69%) and almost rebounded to its 2000 level (73%) in 2002 (72%). Very substantial racial and ethnic disparities in the receipt of this preventive service exist between non-Hispanic blacks and Hispanics relative to non-Hispanic whites. Pneumococcal immunization increased by 2% annually (61%, 63%, and 65%) for the same years. However, very substantial racial and ethnic disparities in the receipt of this preventive service also exist between non-Hispanic blacks, Asians/Pacific Islanders, and Hispanics relative to non-Hispanic whites. CONCLUSIONS: Younger, healthier, elderly persons must be encouraged to receive these immunizations to achieve the 2010 goal of 90% immunization. To reach that goal with no disparities, special efforts will be needed to target racial/ethnic minorities.

Preparing whistleblowers for reporting research misconduct.

Making an allegation of research misconduct can be stressful for a whistleblower. The Research Integrity Officer (RIO) can play an important role in helping reduce the stress by thoroughly discussing what whistleblowers can expect if they make an allegation. Through interviews with 77 RIOs who had recently handled a research misconduct case, we found that RIOs who addressed more topics as well as specific aspects of the topics were more likely to have used some type of memory aide in their initial contact with whistleblowers, talked with ORI staff or other RIOs about "hypothetical" research misconduct cases, or attended a RIO boot camp training. We believe that RIOs who more fully inform whistleblowers are providing timely preparation and building whistleblowers' confidence so they can make a more informed decision about reporting and experience less stress.

Behavioral and socioeconomic correlates of dental problem experience and patterns of health care-seeking.

BACKGROUND: To recognize and better understand barriers to dental care, the authors studied behavioral and socioeconomic factors specific to low-income and minority populations that contribute to patterns of professional health care-seeking. METHODS: A stratified random sample of 27,002 Maryland households participated in a cross-sectional telephone survey. The authors targeted respondents who were low-income non-Hispanic white, non-Hispanic black or Hispanic who had experienced a dental problem or injury during the previous 12 months, as well as similar people with higher incomes for comparison. A total of 94.8 percent (401 of 423) of eligible people completed the survey. RESULTS: Logistic regression modeling (95 percent confidence interval) revealed that the respondents' race or ethnicity and sex were associated with the number of dental problems experienced during the previous 10 years, and that having a regular source of medical care and the type of dental problem experienced were associated with earlier health care-seeking. The first choice of treatment site (physician's office, emergency department or dentist's office) was associated with the respondents' income, pain level, number of problems experienced in the past 10 years and the degree to which their problem was a burden. CONCLUSIONS: The authors found that predisposing, enabling and need-based factors were associated with dental problem experience and patterns of professional health care service use.

Health literacy impact on patient-provider interactions involving the treatment of dental problems.

Health literacy-related problems can interfere with effective doctor-patient communication and effective patient care. This study examined several health literacy-related markers for patients seeking treatment in hospital emergency departments and physician and dentist offices for dental problems and injuries. Participants consisted of low-income white, black, and Hispanic adults who had experienced a dental problem or injury during the previous twelve months and who visited a hospital emergency department, physician, or dentist for treatment. A stratified random sample of Maryland households participated in a cross- sectional telephone survey. Interviews were completed with 94.8 percent of 423 eligible individuals. Multivariable logistic regression analyses were performed. Only 10.0 percent of the respondents expressed a difficulty understanding what they were told by the health provider, while 4.9 percent expressed a difficulty understanding the dental or medical forms they were asked to complete and 6.9 percent reported that they had difficulty getting the health provider to understand their dental problem or injury. Logistic regression analysis found that males and Hispanics were significantly (p<0.05) more likely to experience health literacy-related problems. In general, respondents did not express health literacy-related problems. Additional research is needed to identify health literacy-related barriers to effective patient-provider communication.