Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Multivariate response regression”

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.

At least 127 records · Page 7Linked to original sources
PubMed1990

[Prognostic factors affecting survival and response in patients with advanced non-small cell lung cancer treated with combination chemotherapy].

One-hundred and ninety-nine patients who had an inoperable stage III or IV non-small cell lung cancer (NSCLC) and collected were analyzed on the basis of factors affecting survival duration and response to chemotherapy. These patients were registered into a prospective randomized trial conducted from May of 1986 to April of 1988, and received either cisplatin and vindesine, cisplatin, vindesine and mitomycin C, or cisplatin and etoposide alternating with vindesine and mitomycin C. In the univariate analysis, sex, ECOG's performance status (PS), weight loss within previous 6 months, clinical stage, serum albumin value, serum lactate dehydrogenase level and hemoglobin (Hb) level were considered to be significant factors for survival (p less than 0.05). In the multivariable analysis using Cox's proportional model, clinical stage, PS, sex, weight loss and Hb level were proven to be significant variables for survival in the order of importance. When the response to chemotherapy was included in a conditional multivariable analysis, it was strongly associated with survival duration. A multivariable analysis of response using the logistic regression method demonstrated that female sex, cisplatin, vindesine and mitomycin combination regimen, squamous cell type, and no weight loss were significantly predictive of response outcome. These results are useful when comparing the response data and survival of completed studies and designing future trials in advanced NSCLC.

Adult↗
PubMed2006

Potential predictors of chemotherapy response in ovarian cancer--how do we define chemosensitivity?

OBJECTIVE: The aim of this study was to assess whether microvessel density (measured by CD31), vascular endothelial growth factor (VEGF) or multidrug resistance (MDR1) could determine the response to chemotherapy or act as prognostic factors in ovarian cancer. METHODS: Seventy-nine ovarian specimens were immunostained. Pearson correlation, 1-way ANOVA and chi-square were used for univariate analysis. Kaplan Meier survival curves were used, log-rank was used for univariate analysis and a Cox proportional hazards regression model was used for multivariate evaluation. Response to chemotherapy was assessed after 6 months and again after 1 year. RESULT: Quantifying VEGF proved to be a valuable independent prognostic indicator in progression-free survival (PFS) (p<0.05) and overall survival (OS) (p<0.0001). VEGF correlated with response to chemotherapy at the 6-month interval (r=0.446, p<0.001) but failed to correlate at the 1-year interval. Increased staining with CD31 was associated with decreased PFS (p<0.01) and OS (p<0.01) in univariate but not multivariate analysis. MDR1 failed to act as a prognostic marker or as a predictor of response to chemotherapy. CONCLUSION: VEGF correlates with response to chemotherapy at the 6-month but not the 12-month interval. What should our criteria be for determining sensitivity to chemotherapy? CD31, VEGF and MDR1 do play a role in some ovarian malignancies but other factors are likely to be involved and perhaps molecular profiling will determine which factors will be important for determining the response to chemotherapy.

ATP Binding Cassette Transporter, Subfamily B, Mem↗
PubMed2006

Predictors of response to anti-TNF-alpha therapy among patients with rheumatoid arthritis: results from the British Society for Rheumatology Biologics Register.

BACKGROUND: Anti-tumour necrosis factor-alpha (TNF-alpha) therapies represent an important advancement in therapy for rheumatoid arthritis (RA). However, there remains a proportion of patients who do not improve despite therapy. These drugs are expensive and have the potential of serious toxicity. Therefore, it would be ideal to predict the patients who will respond, so that the use of these drugs can be targetted. OBJECTIVE: To identify the clinical factors present at the start of anti-TNF-alpha therapy that are associated with response at 6 months in patients with RA. METHODS: The British Society for Rheumatology (BSR) Biologics Register collects detailed data on all patients with a rheumatic disease receiving biologic therapy in the UK. We studied all patients with RA who had started etanercept (ETA) or infliximab (INF) and had achieved a minimum 6 months follow-up by 1 October, 2004. The disease status at the baseline and at 6 months was assessed using the Disease Activity Score (DAS28). The response was classified according to the European League against Rheumatism (EULAR) improvement criteria. The effect of baseline characteristics on response was studied using multivariate ordinal logistic regression. RESULTS: 2879 patients were included in this analysis (1267 ETA, 1612 INF). At the start of therapy, the mean age was 55 yrs, disease duration 14 yrs, baseline DAS28 6.7 and health assessment questionnaire (HAQ) 2.1. In all, 28% of ETA and 86% of INF patients were receiving methotrexate. After 6 months, 18% had a good EULAR response, of whom 9% were considered to be in remission and 50% had a moderate response. There was no overall difference in response rate between the two anti-TNF-alpha therapies. A higher baseline HAQ score correlated with a lower response rate while a better response was associated with the current use of NSAIDs and the use of methotrexate (MTX), although the latter only reached statistical significance with ETA [OR 1.82 (95% CI 1.38-2.40)]. There was a lower response rate among current smokers, particularly in patients receiving INF [OR 0.77 (95% CI 0.60-0.99)]. Age, disease duration, rheumatoid factor and the previous number of disease-modifying antirheumatic drugs (DMARDs) did not predict response to either drug. However, females were less likely to achieve remission. CONCLUSIONS: These data support an improved outcome among patients receiving MTX in combination with anti-TNF-alpha therapies. However, the most disabled patients were less likely to respond, despite concurrent MTX. The benefits of NSAIDs may reflect the relative absence of comorbidities in patients who can tolerate these drugs or the continuing presence of reversible inflammatory symptoms. The association of smoking and poor outcome with INF is a novel finding and may reflect alterations in pharmacokinetics. The inability of other baseline disease characteristics to predict the outcome suggests that other factors, including potential genetic differences in drug metabolism, may be influencing the response to anti-TNF-alpha therapies.

Adult↗
PubMedDate not supplied

Predictors of the response to tricyclic antidepressants in major depression.

In a relatively young sample of 108 nonbipolar patients with major depressive disorder and at least moderately severe depression, the contribution of a comprehensive set of socio-demographic, psychosocial, clinical and personality trait variables toward the prediction of short-term, multidimensionally evaluated response to tricyclic antidepressants was investigated by univariate and multivariate analyses. There is a naturalistic study without a standard treatment protocol, but treatment type, duration and dosage were controlled. Only patients who received at least moderate dosages of tricyclic antidepressants for minimum four weeks were included in this study. A relatively low number of predictors emerged. Initial severity of both symptoms and impairment of functioning were the most powerful predictors for all outcome criteria. Illness history was also relevant. Personality traits reflecting vulnerable personality style at the emotional and interpersonal level proved their predictive value. The combination of independent variables by multivariate regression analyses improve the prediction of treatment response.

Adolescent↗
PubMed1998

Ergonomic stressors and upper extremity disorders in vehicle manufacturing: cross sectional exposure-response trends.

OBJECTIVE: To evaluate the association between upper extremity soft tissue disorders and exposure to preventable ergonomic stressors in vehicle manufacturing operations. METHODS: A cross sectional study was conducted in one vehicle stamping plant and one engine assembly plant. A standardised physical examination of the upper extremities was performed on all subjects. An interviewer administered questionnaire obtained data on demographics, work history, musculoskeletal symptoms, non-occupational covariates, and psycho-physical (relative intensity) ratings of ergonomic stressors. The primary exposure score was computed by summing the responses to the psychophysical exposure items. Multivariate regression analysis was used to model the prevalence of disorders of the shoulders or upper arms, wrists or hands, and all upper extremity regions (each defined both by symptoms and by physical examination plus symptoms) as a function of exposure quartile. RESULTS: A total of 1315 workers (85% of the target population) was examined. The prevalence of symptom disorders was 22% for the wrists or hands and 15% for the shoulders or upper arms; cases defined on the basis of a physical examination were about 80% as frequent. Disorders of the upper extremities, shoulders, and wrists or hands all increased markedly with exposure score, after adjustment for plant, acute injury, sex, body mass index, systemic disease, and seniority. CONCLUSIONS: Musculoskeletal disorders of the upper extremities were strongly associated with exposure to combined ergonomic stressors. The exposure-response trend was very similar for symptom cases and for physical examination cases. It is important to evaluate all dimensions of ergonomic exposure in epidemiological studies, as exposures often occur in combination in actual workplaces.

Adult↗
PubMed2004

Vascular response to angiotensin II predicts long-term prognosis in patients undergoing coronary artery bypass grafting.

Persistent activation of the renin-angiotensin system leads to downregulation of the angiotensin type-1 receptor, and consequently, to a decreased response to exogenous angiotensin II. In the present study, we investigated the association of angiotensin II responsiveness to clinical outcome after coronary artery bypass grafting (CABG). We studied the responsiveness to exogenous angiotensin II in human thoracic artery preparations of 114 CABG patients. Mean duration of follow-up was 7.3+/-0.1 years, during which 21 patients experienced a cardiovascular event. A diminished response to angiotensin II remained in multivariate Cox regression analysis, after adjustment for sex, age, blood pressure, and number of diseased coronary arteries, the strongest predictor for cardiovascular events (relative risk, 3.37 [95% confidence interval, 1.20 to 9.51]; P=0.022). Furthermore, diminished response to angiotensin II was associated with an increased mean arterial pressure (102.85+/-1.38 versus 97.40+/-1.37; P=0.003) and a nonsignificant increase in angiotensin-converting enzyme activity, suggestive for a persistently activated renin-angiotensin system. In conclusion, these results suggest that in patients undergoing CABG, a diminished vascular responsiveness of the thoracic artery to exogenous angiotensin II is related to an increased risk of future cardiovascular events.

Aged↗
PubMed2000

Stepwise multimodal approach in the treatment of Kasabach-Merritt syndrome.

BACKGROUND: The purpose of the present study was to evalutate the various treatment modalities for Kasabach-Merritt syndrome (KMS) and to identify the most reliable treatment modalities. METHODS: A retrospective analysis was performed on 37 KMS patients who were admitted to Seoul National University Hospital between January 1979 and June 1999. Age, sex, locations of the hemangiomas, clinical symptoms and hematologic data were analyzed by multivariate logistical regression analysis to determine the response to various treatment modalities. RESULTS: Twenty-four of 37 patients (20 boys and 17 girls) were diagnosed during infancy. The locations of hemangioma were superficial skin in 31 patients, combined skin and visceral organs in four patients and visceral organs only in two patients. Tumor size was more than 5 cm in diameter in all cases, except for one. The treatment principle of KMS in our center involves a stepwise multimodal approach. Of the 37 cases, surgical resection was performed in two. Steroids were tried initially in 35 patients over a 1 week period, with an initial response rate of 11.4% (4/35). Combined steroid/radiation treatment was given in 28 cases, with a response rate of 75.0% (21/28). Four of five patients with additional interferon (IFN)-alpha therapy improved. No significant correlations were found between the prognostic factors, such as age, sex, size and locations of the hemangioma, hematologic data and the treatment modalities. CONCLUSIONS: Based on the experiences in a single center over a period of 20 years, we recommend a stepwise multimodal approach for the treatment of KMS; for example, steroid, radiation therapy and IFN-alpha in this order. However, surgical treatment is helpful if total resection is possible.

Abnormalities, Multiple↗
PubMed1990

Cervical mucus and serum estradiol as predictors of response to progestin challenge.

The present study was undertaken to assess the correlation between and relative predictive value of each of the following variables and progestin-induced withdrawal bleeding: cervical mucus appearance, serum E2 level, patient age, duration of amenorrhea, smoking and exercise habits, and body composition. Of 120 oligomenorrheic and amenorrheic women evaluated, only cervical mucus appearance and serum E2 level were significantly associated with response to progestin challenge. A multivariate logistical regression analysis showed cervical mucus to be the most predictive variable followed by serum E2 level. No absolute E2 level was found to discriminate between those who did and those who did not have withdrawal bleeding after progestin challenge. These data suggest that office examination of cervical mucus may be a useful indicator and guideline in planning therapy.

Administration, Oral↗
PubMed1998

Lithium maintenance treatment of depression and mania in bipolar I and bipolar II disorders.

OBJECTIVE: Effects of long-term lithium treatment for depressive and manic phases of type I and type II bipolar disorders were compared. METHOD: Clinical research records of 317 patients with DSM-IV-defined bipolar disorder (188 with type I and 129 with type II) were analyzed for frequency and duration of affective episodes and hospitalizations before (mean = 8.38 years) versus during (mean = 6.35 years) lithium maintenance treatment. Treatment effects were also assessed by survival analysis of interepisode intervals and by multivariate regression testing for factors associated with response to treatment. RESULTS: Bipolar I and bipolar II patients were ill before treatment a similar percentage of time, but the subtype distinction was supported descriptively. Lithium had superior benefits in type II patients, with significantly greater reduction of episodes per year and of the percentage of time ill. Reduction of depressive morbidity was similarly strong in both diagnostic types. During treatment, bipolar II patients had 5.9-fold longer interepisode intervals and were twice as likely as type I patients to have no new episodes. Starting lithium maintenance earlier predicted greater improvement. CONCLUSIONS: Lithium maintenance yielded striking long-term reductions of depressive as well as manic morbidity in both bipolar disorder subtypes, with greater overall benefits in type II patients and with earlier treatment.

Adult↗
PubMedDate not supplied

Lattice theory analysis of the relationship of hyaline membrane disease and fetal pneumonia in 96 perinatal autopsies.

A previous study of hyaline membrane disease of the newborn demonstrated that the hyaline membrane was derived from necrotic epithelial cells, and it was suggested that a no flow-reflow episode affecting the pulmonary circulation might account for the cell necrosis. An unexpected statistical finding in that study was a highly significant negative correlation between hyaline membrane disease and chorioamnionitis. Because it seemed to us that fetal infection could be expected to produce the circulatory pathophysiology suggested to produce hyaline membrane disease, we reexamined the issue by reviewing a group of autopsied infants. Data were collected on clinicopathologic variables from 42 stillborn and 54 liveborn infants less than or equal to 12 h of age with lung and placenta slides available for review. Correlations and multivariate regression analysis showed that the inflammatory responses in membranes, cord, and lung are interrelated but that hyaline membrane disease develops independently. However, lattice theory analysis, which represents the pathogenesis of a progressive disease process as edges along a mathematical lattice or hyperdimensional cube, showed that this separation was not absolute. Despite the negative correlation of hyaline membrane disease and fetal pneumonia, there were 10 (10%) patients with both conditions and a total of 19 (20%) with hyaline membrane disease and inflammation of membranes, cord, and/or lung. We suggest that the coexistence of hyaline membrane disease and fetal pneumonia in some patients implies a possible pathogenetic relationship between the two entities. Given the nature of the two processes, it seems likely that the direction of causality is that fetal pneumonia would give rise to the hyaline membrane disease of the newborn.

Autopsy↗
PubMed2003

[A case-control study on the risk factors of injuries in middle school students in southern part of Anhui province].

OBJECTIVE: To investigate the risk factors of injuries in middle school students and to provide basis for the prevention and reduction of such incidence. METHODS: 1:1 case-control study was conducted on 254 cases and 254 controls on a basis of sex, age and grade. The relationship between risk factors and injuries were analyzed by conditional univariate and multivariate logistic regression. RESULTS: Seven risk factors responsible for the incidence of injury were identified as follows: high risk behaviors (OR = 18.0600), negligence of defence (OR = 12.6455), scramble (OR = 9.6552), father being illiterate (OR = 7.7191), risky environment around their houses (OR = 5.7402), extrovert temperament (OR = 5.4707) and mother being illiterate (OR = 3.0581). We also distinguished 5 protective factors as follows: education on safety (OR = 0.2356), harmonic relation between parents (OR = 0.4941), one-child per families (OR = 0.5233), students were more knowledgeable and having positive attitude towards road traffic (OR = 0.5340) and high economy level (OR = 0.5609). CONCLUSION: The injuries in middle school students were caused by multiple factors, hence should carry out intervention measures to modify the influencing factors of injuries. Injury prevention and control program should focus on certain strategies such as education and supervision of environment hazards.

Accident Prevention↗
PubMed2006

Adverse childhood experiences and lifetime suicide ideation: a cross-sectional study in a non-psychiatric hospital setting.

We have assessed the effect of adverse childhood experiences on the lifetime prevalence of suicide ideation in a cross-sectional study involving 182 patients aged 18 to 44 years, consecutive attenders at an A&E review clinic. All participants were interviewed by a psychologist using standardised questionnaire instruments addressing participants' demographic characteristics, drug use, depressed mood, eight major categories of adverse childhood experiences (including physical, emotional and sexual abuse) and suicide ideation. The response rate was 73%. In multivariate logistic regression analyses, those with a history of two or more forms of childhood adversity relative to those with none were at increased risk of depressed mood (OR = 5.5, 95% CI = 2.3-13.3) and suicide ideation (OR = 3.5, 95% CI = 1.5-8.3). The findings emphasise the need to set suicide prevention within the broader context of society's obligation to protect children from physical, emotional and sexual abuse.

Adolescent↗
PubMed1997

Nonsurgical factors predictive of postoperative hearing for patients with vestibular schwannoma.

OBJECTIVE: The purposes of the study were to determine whether preoperative cochlear reserve as measured by evoked otoacoustic emissions (EOAE) as well as other hearing variables often associated with hearing preservation are correlated with hearing preservation after tumor removal and to determine whether any hearing variables are independent of tumor size as a predictor of hearing preservation. STUDY DESIGN: Preoperative audiologic data for 104 patients having vestibular nerve schwannomas removed via a retrosigmoid surgical approach were reviewed and subjected to factor analysis. SETTING: All patients were seen at the Mayo Clinic, Rochester, Minnesota. PATIENTS: The patient sample was divided into two groups based on hearing thresholds after surgery. Group I consisted of 73 ears without hearing preservation. The remaining 31 ears, group II, had preserved hearing (defined as average postoperative pure-tone thresholds < or = 85-dB HL for 0.5, 1, 2, and 3 kHz). MAIN OUTCOME MEASURES: Variables not predictive of hearing preservation were age, gender, tumor laterality, and cochlear reserve (EOAE). Variables predictive of hearing preservation were small tumor size, pure-tone hearing sensitivity, speech reception thresholds, word recognition scores, integrity of cochlear nerve (acoustic reflex thresholds, and auditory brain stem response [ABR] waveforms). RESULTS: A multivariate logistic regression analysis showed that only word recognition scores at 40-dB sensation level were independent of tumor size as a predictor of hearing preservation.

Acoustic Stimulation↗
PubMed2000

Human immunodeficiency virus-infected persons with mutations conferring resistance to zidovudine show reduced virologic responses to hydroxyurea and stavudine-lamivudine.

The baseline predictors of poor virologic response (<0.5 log decrease in plasma virus load) were examined in two 1996 pilot trials of combination nucleoside-analogue therapy. One trial examined the addition of hydroxyurea to didanosine therapy; the other examined stavudine-lamivudine in combination. In both, predictors of virologic response included the presence of mutations associated with zidovudine resistance. For hydroxyurea, the odds ratio (OR) of failure to achieve a short-term (4 weeks) virologic response in a bivariate logistic regression model was 30.8 (95% confidence interval [CI], 1.75-543; P=.02) for use of lower dose hydroxyurea (500 mg/day) and 14.7 (95% CI, 1.1-200; P=.04) for the presence of a zidovudine-related mutation. For the stavudine-lamivudine study, the OR of failure to achieve a virologic response at 4 weeks in a multivariate logistic regression model was 23 (95% CI, 2.7-199; P=.004) for the presence of a mutation at codon 215.

Anti-HIV Agents↗
PubMed2006

A mixed-effects regression model for longitudinal multivariate ordinal data.

A mixed-effects item response theory model that allows for three-level multivariate ordinal outcomes and accommodates multiple random subject effects is proposed for analysis of multivariate ordinal outcomes in longitudinal studies. This model allows for the estimation of different item factor loadings (item discrimination parameters) for the multiple outcomes. The covariates in the model do not have to follow the proportional odds assumption and can be at any level. Assuming either a probit or logistic response function, maximum marginal likelihood estimation is proposed utilizing multidimensional Gauss-Hermite quadrature for integration of the random effects. An iterative Fisher scoring solution, which provides standard errors for all model parameters, is used. An analysis of a longitudinal substance use data set, where four items of substance use behavior (cigarette use, alcohol use, marijuana use, and getting drunk or high) are repeatedly measured over time, is used to illustrate application of the proposed model.

Data Interpretation, Statistical↗
PubMed2000

Predictive factors of response to cisplatin-based chemotherapy and the relation of response to survival in patients with metastatic urothelial cancer.

PURPOSE: To identify pretreatment variables predicting overall and complete response to cisplatin-based chemotherapy for metastatic urothelial cancer, and to study the relation between response and the duration of survival. PATIENTS AND METHODS: A total of 119 evaluable patients with recurrent locally advanced or metastatic urothelial cancer received cisplatin-based combination chemotherapy in four consecutive phase II studies from 1987 to 1997. The relationship of pretreatment variables and response was evaluated with logistic regression, and prognostic factors for survival were analyzed with Cox's multivariate model. RESULTS: Response was achieved in 49% of the patients with a complete response rate of 15%. Good performance status and absence of bone metastases were independently predictive of overall response. Good performance status and normal hemoglobin were independently predictive of complete response. Median survival was 8.9 months. Performance status, alkaline phosphatase, s-creatinine, liver and bone metastases were independent prognostic factors for survival. Median survival was 12.4 months in responding patients and 6.3 in nonresponding patients. Response to chemotherapy was included in the multivariate model and was the strongest prognostic factor for survival. CONCLUSION: The presence of bone metastases, low hemoglobin or poor performance status predicts decreased chance of response to chemotherapy. Response to chemotherapy is an independent prognostic factor for prolonged survival in patients with metastatic urothelial cancer.

Aged↗
PubMed1994

Delayed-type hypersensitivity anergy in human immunodeficiency virus-infected persons screened for infection with Mycobacterium tuberculosis.

A total of 479 human immunodeficiency virus (HIV)-infected persons at an HIV clinic in Florida and a tuberculosis clinic in New Jersey were skin-tested with tuberculin, tetanus toxoid, mumps antigen, and Candida antigen in a study of the prevalence of delayed-type hypersensitivity (DTH) anergy and the usefulness of two-step tuberculin testing in this population. Of the patients tested, 12% had a positive (> or = 5-mm) response to tuberculin; 57%, 45%, and 35% had a positive (> or = 3-mm) response to Candida antigen, tetanus toxoid, and mumps antigen, respectively; and 31% were anergic (< 3 mm of induration in response to each antigen). In a multivariate logistic regression model, anergy was significantly associated with a history of Kaposi's sarcoma, Pneumocystis carinii pneumonia, or oral candidiasis and with White race. Anergy was four times and 15 times as likely for persons with CD4+ T-lymphocyte counts of 200-400/mm3 and < 200/mm3, respectively, as for persons with > 499 CD4+ T lymphocytes/mm3. Of 103 patients who were tuberculin-tested a second time after their initial test result was negative, seven had > or = 5 mm of induration in response to the second test; only one of these patients was anergic at the initial screening. The findings of this study indicate that DTH antigens should be used in conjunction with tuberculin testing and that two-step tuberculin testing is not an alternative to anergy testing but may be useful for the detection of infection with Mycobacterium tuberculosis in nonanergic HIV-infected patients.

AIDS-Related Opportunistic Infections↗
PubMed1999

Similarity as a risk factor in drug-name confusion errors: the look-alike (orthographic) and sound-alike (phonetic) model.

BACKGROUND: One of every four medication errors reported in the United States is a name-confusion error. The rate of name-confusion errors might be reduced if new and confusing names were not allowed on the market and if safeguards could be put in place to avoid confusion between existing names. OBJECTIVES: To evaluate several prognostic tests of drug-name confusion, alone and in combination, with respect to their sensitivity, specificity, and overall accuracy. RESEARCH DESIGN: Case-control study. Twenty-two different computerized measures of orthographic similarity, orthographic distance, and phonetic similarity were used to compute similarity/distance scores for n = 1,127 cases (ie, pairs of names that appeared in published error reports or national error databases) and n = 1,127 controls. MAIN OUTCOME MEASURES: Mean similarity/distance scores were compared across cases and controls. The performance of each measure at distinguishing between cases and controls was evaluated by tenfold crossvalidation. Dose-response relationships were examined. Univariate and multivariate logistic regression models were formed and evaluated by 10 fold crossvalidation. RESULTS: Cases had significantly higher similarity scores than controls. Every measure of similarity proved to be a significant risk factor for error. There was a significant increasing trend in the odds-ratio as a function of similarity. A three-predictor logistic regression model had crossvalidated sensitivity of 93.7%, specificity of 95.9% and accuracy of 94.8%. CONCLUSIONS: A sensitive and specific test of drug-name confusion potential can be formed using objective measures of orthographic similarity, orthographic distance, and phonetic distance.

Analysis of Variance↗