Baser, O., Gardiner, J.C., Bradley, C.J., and others (2006). "Longitudinal analysis of censored medical cost data." (AHRQ grant HS14206). Health Economics 15, pp. 513-525.
Challenges in analyzing medical cost data include addressing askewness in cost distributions, heterogeneity across samples, and complexities due to censoring. This paper applies the inverse probability weighted (IPW) least-squares method to estimate the effects of lung cancer treatment on total medical cost for 201 Medicare patients, subject to censoring, in a panel-data setting. The researchers used IPW pooled ordinary-least squares (POLS) and IPW random effects (RE) models. Because total medical cost might not be independent of survival time under administrative censoring, unweighted POLS and RE cannot be used with censored data to assess the effects of certain explanatory variables. Even under the violation of this independency, IPW estimation yielded consistent asymptotic normal coefficients with easily computable standard errors.
Bisson, G.P., Gross, R., Strom, J.B., and others (2006). "Diagnostic accuracy of CD4 cell count increase for virologic response after initiating highly active antiretroviral therapy." (AHRQ grant HS10399). AIDS 20(12), pp. 1613-1619.
Monitoring plasma HIV-1 RNA levels (viral loads) is critical to identifying treatment failure in patients taking highly active antiretroviral therapy (HAART). HIV treatment programs in resource-constrained settings could consider the use of CD4 cell count increases to triage patients for viral load testing. However, the authors of this study call for more accurate approaches to monitoring virologic failure. They found that an increase in CD4 cell count after beginning HAART had only moderate discriminating ability to identify patients with an undetectable viral load (400 copies/ml or less). The predictive ability of CD4 cell count was even lower in patients with lower baseline CD4 cell counts. Their findings were based on a retrospective study at two HIV care clinics in Gaborone, Botswana.
Gardiner, J.C., Luo, Z., Bradley, C.J., and others (2006). "A dynamic model for estimating changes in health status and costs." (AHRQ grant HS14206). Statistics in Medicine 25, pp. 3648-3667.)
These authors describe development of an innovative method to estimate changes in health status and costs. They used a Markov model to estimate the transition probabilities between health states and the impact of patient variables on transition intensities for 624 cancer patients. They used a mixed-effects model for sojourn costs, with transition times as random effects and patient variables as fixed effects. They combined the models to estimate net present values of expenditures, as a function of patient characteristics, by cancer site and cancer stage over a 2-year period.
Haidet, P., Kroll, T.L., and Sharf, B.F. (2006). "The complexity of patient participation: Lessons learned from patients' illness narratives." (AHRQ grant HS10876). Patient Education and Counseling 62, pp. 323-329.
Based on interviews with primary care patients in Texas, this paper provides insight into the patient's perspective of what active participation in care means. Patients' illness narratives reflected several themes related to patient participation. For example, they described how central their illness was in their overall life story and how changeable they believed their illness to be. They also described actions they pursued in the context of these illness narratives and the role of partnership with their physician in health decision making and illness management. Generally, patients' illness management strategies were explained by these four themes in dynamic interplay, with unique variations for each individual. The researchers call for more studies to explore how these themes influence communication between patients and physicians.
Harless, D.W., and Mark, B.A. (2006, October). "Addressing measurement error bias in nurse staffing research." (AHRQ grant HS10153). HSR: Health Services Research 41(5), pp. 2006-2024.
Nurse staffing allocation methods induce substantial attenuation bias, the authors conclude; however, there are easily implemented estimation methods that overcome this bias. They analyzed data from the California Office of Statewide Health Planning and Development to estimate the measurement error and resulting bias from applying different methods to allocate nursing staff. The bias induced by the adjusted patient days method was smaller than for other methods, but the bias was still substantial. Instrumental variable estimation, using one staffing allocation measure as an instrument for another, addressed this bias. However, only particular choices of staffing allocation measures and instruments are suitable.
Hays, R.D., Brown, J., Brown, L.U., and others (2006, November). "Classical test theory and item response theory analyses of multi-item scales assessing parents' perceptions of their children's dental care." (AHRQ grant HS00924). Medical Care 44(11), S60-S68.
Evaluation of a beta (pre-release) version of the Consumer Assessment of Healthcare Providers and Systems (CAHPS) dental care survey found that it performed well. The beta version included two global rating items (dental care, dental plan), as well as multi-item scales that assessed getting needed dental care, getting care quickly, communication with dental providers, office staff, and customer services. The researchers examined 2001 and 2002 survey data of families with children between ages 4 to 18, who were enrolled in one of five dental plans for a year or longer. Item missing data rates were low. Item-scale correlations for hypothesized scales (corrected for overlap) tended to exceed correlations of items with other scales. Classical test theory analyses identified 5 of 10 communication items that did not perform well. Item response theory painted a more promising picture than classical test theory for the two communication items that assessed access to an interpreter when needed.
Hermann, R.C., Chan, J.A., Provost, S.E., and Chiu, W.T. (2006, October). "Statistical benchmarks for process measures of quality of care for mental and substance use disorders." (AHRQ grant HS10303). Psychiatric Services 57(10), pp. 1461-1467.
Benchmarks can be used to set achievable goals for improving care; yet, until now, they have not been available for mental health care. This article describes the application of a method for developing statistical benchmarks for 12 process measures of quality of care for mental and substance use disorders. Three measures involved antidepressant treatment, two involved antipsychotic treatment, and one involved mood stabilizers for bipolar disorder. Six other measures involved followup treatment visits. Benchmarks for provider-level performance ranged from 59.7 percent to 97.7 percent, markedly higher than the mean results, which ranged from 9.4 percent to 65.4 percent. The researchers conclude that statistical benchmarks can be applied to results from quality assessment of mental health care.
Katerndahl, D. and Crabtree, B. (2006, September). "Creating innovative research designs: The 10-year methodological think tank case study." (AHRQ grant HS08775). Annals of Family Medicine 4(5), pp. 443-449.
Important but complex research issues have emerged that defy direct application of most available research designs and methods in which investigators have been trained. Each year, a group of three to four methodologists with expertise balanced between quantitative and qualitative backgrounds is invited to the annual Methodological Think Tank, which has been held in conjunction with the Primary Care Research Methods and Statistics Conference in San Antonio since 1994. Over 2 days, participants discuss a research question selected from those submitted in response to a call for proposals. During the first half-day, the experts explore the content area with the investigator, often challenging beliefs and assumptions. During the second half-day, the think tank participants systematically prune potential approaches until a desirable research method is identified. To date, the most recent seven think tanks have produced fundable research designs.
MacAdam, H., Zaoutis, T.E., Gasink, L.B., and others (2006). "Investigating the association between antibiotic use and antibiotic resistance: Impact of different methods of categorizing prior antibiotic use." (AHRQ grant HS10399). International Journal of Antimicrobial Agents 28, pp. 325-332.
Prior antibiotic use has been identified as one of the most consistent and modifiable risk factors associated with antibiotic-resistant infections. However, this review of studies concludes there has been no consistent approach to categorizing prior antibiotic use in studies of risk factors for extended-spectrum-B-lactamase-producing Escherichia coli and Klepsiella spp. (ESBL-EK). Among the 20 studies reviewed, there was tremendous variability in how prior antibiotic use was categorized—for example, by agent, class, spectrum, and/or a combination of these. Yet different categorization schemes had a substantial impact on the antibiotic exposures associated with antibiotic-resistant infections.
McHorney, C.A. and Fleishman, J.A. (2006, November). "Assessing and understanding measurement equivalence in health outcome measures." Medical Care 44(11), pp. S205-S210.
This article provides an overview of this journal's special supplement on issues of measurement equivalence in diverse populations, particularly populations characterized by health disparities. Measurement bias can lead to flawed population forecasts for service needs or resource allocations, as well as misguided research on health disparities. Measurement bias could also degrade the validity, interpretability, and generalizability of "real-world" outcome effectiveness studies, note the authors. Articles in the supplement provide an excellent overview of the importance of considering differential item functioning (DIF) when making group comparisons, as well as different techniques for identifying DIF, to avoid measurement bias when measuring health outcomes.
Meterko, M., Young, G.J., White, B., and others (2006, October). "Provider attitudes toward pay-for-performance programs: Development and validation of a measurement instrument." (AHRQ grant HS13591). HSR: Health Services Research 41(5), pp. 1959-1978.
These researchers developed a questionnaire for assessing physician attitudes toward pay-for-performance incentive programs, based on an extensive review of the literature and discussions with experts in the field. They distributed a revised version to 2,497 primary care physicians affiliated with 2 of 7 demonstration sites that are testing different ways to give providers incentive to improve quality of care. Each of several attitudinal measures—awareness and understanding of incentives, clinical relevance, cooperation, unintended consequences, control, financial salience, and impact—was a significant predictor of a provider's perceived impact of quality-based financial incentives. The researchers conclude that it is possible to identify and measure the key salient features of pay-for-performance programs using a valid and reliable 26-item survey.
Nace, G.S., Grumlich, J.F., and Aldag, J.C. (2006). "Software design to facilitate information transfer at hospital discharge." (AHRQ grant HS15084). Informatics in Primary Care 14, pp. 109-119.
This paper describes development of discharge software to overcome communication barriers. The authors found that the discharge software can help inpatient physicians transfer timely, complete, and legible information to outpatient physicians, pharmacists, and patients. Use of the software revealed that physician factors significantly affected the time to complete a discharge while using the software. For example, an increased number of accesses (log-ins) and more free text typing lengthened the time to complete the computerized discharge. Patient-related factors that increased physician time were discharge diagnoses, prescriptions, and length of hospital stay.
Napoles-Springer, A.M., Santoyo-Olsson, J., O'Brien, H., and Steward, A.L. (2006, November). "Using cognitive interviews to develop surveys in diverse populations." (AHRQ grant HS10599); and Napoles-Springer, A.M., and Stewart, A.L. (2006, November). "Overview of qualitative methods in research with diverse populations." (HS10856). Medical Care 44(11), pp. S21-S30, S5-S9.
Cognitive interviews are used widely in questionnaire development to detect items that are not understood by respondents as intended by the survey developers. In the first paper, the authors describe an interaction analysis approach using qualitative data analysis software to analyze transcripts of cognitive interviews. Their goal was to develop a survey to assess the quality of interpersonal processes of care for diverse patients. They completed interviews with 48 Latino, black, and white adults. They identified problems with 126 of 159 survey items (79 percent). Behavior coding identified 32 problematic items (20 percent). Interaction analysis of the survey transcript and retrospective probes identified 94 additional problematic items (59 percent). In the second paper, the authors provide an overview of qualitative methods in research with diverse populations.
Raab, S.S. (2006, May). "Improving patient safety through quality assurance." (AHRQ grant HS13321). Archives of Pathology & Laboratory Medicine 130, pp. 633-637.
Anatomic pathology laboratories use several quality assurance tools to detect errors and to improve patient safety. The authors of this paper reviewed several pathology laboratory patient safety quality assurance practices and found that anatomic pathology error frequencies varied according to the detection method used. In addition, there was a lack of standardization across laboratories, even for governmentally mandated quality assurance practices such as cytologic-histologic correlation. Nevertheless, pathology laboratories are starting to use data from quality assurance practices for initiatives to reduce pathology errors.
Robertson, K.B., Janssen, W.J., Saint, S., and Weinberger, S.E. (2006, November). "The missing piece." (AHRQ grant HS11540). The New England Journal of Medicine 355(18), pp. 1913-1918.
This paper describes the case of a young man who contracted a food-borne parasitic infection during his work with raw fish as a sushi chef. The man had abdominal pain and respiratory cough, sputum, and dizziness on exertion. The clinician focused on the sources of abdominal pain and respiratory disease to rule out certain possibilities. As more information became available, he was able to identify more specific patterns, including pleuroparenchymal lung disease and subcutaneous nodules. The patient's striking peripheral-blood eosinophilia (high white cell count) proved to be the sentinel clue, pointing to a parasitic infection with gastrointestinal and pulmonary involvement as the likely diagnosis. The final diagnosis was paragonimiasis, a food-borne parasitic infection caused by numerous species of lung flukes. Humans are infected when they ingest raw or partially cooked crabs or crayfish containing paragonimus metacercariae.
Schackman, B.R., Gebo, K.A., Walensky, R.P., and others (2006, November). "The lifetime cost of current human immunodeficiency virus care in the United States." (AHRQ Contract No. 290-01-0012). Medical Care 44(11), pp. 990-997.
This study calculated that from the time of entering HIV care in 2001, a person with HIV disease could expect to live another 24.2 years, compared with 4 years in 1997. The lifetime HIV care costs (2004 dollars) would be $618,900 (undiscounted) for adults who begin antiretroviral therapy (ART) with a CD4 cell count less than 350. Nearly three-fourths (73 percent) of the cost would be for antiretroviral medications, 13 percent for inpatient care, 9 percent for outpatient care, and 5 percent for other HIV-related medications and laboratory costs.
The total lifetime care cost would be comparable to the estimated $599,000 (undiscounted) lifetime medical cost for nonelderly women with cardiovascular disease, who also are expected to live long with appropriate medical management. Life expectancy and costs are slightly lower for patients who begin ART with a CD4 cell count less than 200 (an indicator of advanced disease, and the point at which opportunistic infections become a problem).
Sexton, J.B., Makary, M.A., Tersigni, A.R., and others (2006, November). "Teamwork in the operating room." (AHRQ grants HS14246 and HS11544). Anesthesiology 105(5), pp. 877-884.
It is possible to use the teamwork climate domain of the Safety Attitudes Questionnaire to assess teamwork in the hospital operating room (OR), concludes this study. The researchers surveyed OR personnel in 60 U.S. hospitals about the survey's 6 teamwork items. These included difficulty speaking up, conflict resolution, physician-nurse collaboration, feeling supported by others, asking questions, and heeding nurse input. The researchers grouped individual-level responses to a single score at each hospital OR level and used multivariate analysis of items and scale to detect differences at the hospital OR level and by caregiver type. They found that teamwork climate differed significantly by hospital and OR caregiver type. This tool and initial benchmarks should allow hospitals to compare their OR teamwork climate to national means, in an effort to focus more on what excellent surgical teams do well.
Return to Contents
AHRQ Publication No. 07-0018
Current as of February 2007
Internet Citation:
Research Activities Newsletter. February 2007, No. 318. AHRQ Publication No. 07-0018. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/research/feb07/
540 Gaither Road Rockville, MD 20850