Delayed medical care after diagnosis in a US national probability sample of persons infected with human immunodeficiency virus.

OBJECTIVE: To identify health care and patient factors associated with delayed initial medical care for human immunodeficiency virus (HIV) infection. DESIGN: Survey of a national probability sample of persons with HIV in care. SETTING: Medical practices in the contiguous United States. PATIENTS: Cohort A (N = 1540) was diagnosed by February 1993 and was in care within 3 years; cohort B (N = 1960) was diagnosed by February 1995 and was in care within 1 year of diagnosis. MAIN OUTCOME MEASURE: More than 3- or 6-month delay. RESULTS: Delay of more than 3 months occurred for 29% of cohort A (median, 1 year) and 17% of cohort B. Having a usual source of care at diagnosis reduced delay, with adjusted odds ratios (ORs) of 0.61 (95% confidence interval [CI], 0.48-0.77) in cohort A and 0. 70 (95% CI, 0.50-0.99) in cohort B. Medicaid coverage at diagnosis showed lower adjusted ORs of delay compared with private insurance (cohort A: adjusted OR, 0.52; 95% CI, 0.30-0.92; cohort B: adjusted OR, 0.48; 95% CI, 0.27-0.85). Compared with whites, Latinos had 53% and 95% higher adjusted ORs of delay (P<.05) in cohorts A and B, respectively, and African Americans had a higher adjusted OR in cohort A (1.56; 95% CI, 1.19-2.04). The health care factors showed similar effects on delay of greater than 6 months. CONCLUSIONS: Medicaid insurance and a usual source of care were protective against delay after HIV diagnosis. After full adjustment, delay was still greater for Latinos and, to a lesser extent, African Americans compared with whites.

The effect of hospital experience on mortality among patients hospitalized with acquired immunodeficiency syndrome in California.

BACKGROUND: Several studies in the 1980s suggested that mortality rates for patients hospitalized with the acquired immunodeficiency syndrome (AIDS) were lower in hospitals that cared for greater numbers of AIDS patients. We sought to determine whether this observation persisted in the mid-1990s in California. SUBJECTS AND METHODS: We performed an analysis of hospital discharge data for 7,901 adults discharged with human immunodeficiency virus (HIV) or AIDS-related diagnoses from all acute care hospitals (n = 333) in California during 1994. The main outcome measure was in-hospital mortality, adjusted for severity of illness, comorbidity, prior hospitalizations, and other patient and hospital characteristics. RESULTS: Among 7,901 persons hospitalized with AIDS, the unadjusted inpatient mortality was 9.0%. The adjusted mortality rate varied significantly (P <0.0001) from 12.4% among institutions with the lowest quartile of AIDS experience to 10.3%, 6.3%, and 7.6% by quartile of greater AIDS experience. Increasing severity of illness, comorbidity, and previous hospitalizations were also significant predictors of increased mortality. Sex, race, and insurance status were not associated with hospital mortality. CONCLUSIONS: Hospital mortality for AIDS patients was greater among less experienced hospitals. The difference in mortality rate was equivalent to more than four additional deaths per 100 patients with AIDS admitted to the least experienced hospitals. This finding was not explained by severity of illness, comorbidity, or other case-mix variables.

Constitutional symptoms and health-related quality of life in patients with symptomatic HIV disease.

PURPOSE: To assess the severity of constitutional symptoms in persons with human immunodeficiency virus (HIV) infection, and their relationship to health-related quality of life (HRQOL). PATIENTS AND METHODS: Two hundred five HIV-infected patients (93% male, 26% African American, 28% Latino, 39% white, 7% other ethnicity) with diarrhea, fever, or weight loss were studied at a county hospital and a Veterans Administration hospital in southern California. Consenting subjects were administered a battery that included 11 scales measuring various aspects of health-related quality of life and detailed questions about six constitutional symptoms or symptom complexes (myalgias, exhaustion, anorexia/nausea/vomiting, night sweats, fever, and weight loss) as well as about other manifestations of HIV disease. RESULTS: Constitutional symptoms except weight loss were all strongly related to all measures of quality of life. On 0 (worst) to 100 (best) point scales, mean scores ranged from 34 (for individuals having all five symptoms other than weight loss) to 78 (for those with none) for physical function, 43 to 79 for emotional well-being, and 36 to 73 for social function. Adjustment for helper T-lymphocyte counts, duration of illness, and demographic characteristics did not diminish these associations. CONCLUSION: The presence, number, and severity of constitutional symptoms in HIV disease is strongly related to health-related quality of life in symptomatic HIV-infected individuals. Identifying and treating these very common symptoms has the potential to improve quality of life in these patients.

Health-related quality of life in patients with human immunodeficiency virus infection in the United States: results from the HIV Cost and Services Utilization Study.

PURPOSE: To measure health-related quality of life among adult patients with human immunodeficiency virus (HIV) disease; to compare the health-related quality of life of adults with HIV with that of the general population and with patients with other chronic conditions; and to determine the associations of demographic variables and disease severity with health-related quality of life. SUBJECTS AND METHODS: We studied 2,864 HIV-infected adults participating in the HIV Cost and Services Utilization Study, a probability sample of adults with HIV receiving health care in the contiguous United States (excluding military hospitals, prisons, or emergency rooms). A battery of 28 items covering eight domains of health (physical functioning, emotional well-being, role functioning, pain, general health perceptions, social functioning, energy, disability days) was administered. The eight domains were combined into physical and mental health summary scores. SF-36 physical functioning and emotional well-being scales were compared with the US general population and patients with other chronic diseases on a 0 to 100 scale. RESULTS: Physical functioning was about the same for adults with asymptomatic HIV disease as for the US population [mean (+/- SD) of 92+/-16 versus 90+/-17) but was much worse for those with symptomatic HIV disease (76+/-28) or who met criteria for the acquired immunodeficiency syndrome (AIDS; 58+/-31). Patients with AIDS had worse physical functioning than those with other chronic diseases (epilepsy, gastroesophageal reflux disease, clinically localized prostate cancer, clinical depression, diabetes) for which comparable data were available. Emotional well-being was comparable among patients with various stages of HIV disease (asymptomatic, 62+/-9; symptomatic, 59+/-11; AIDS, 59+/-11), but was significantly worse than the general population and patients with other chronic diseases except depression. In multivariate analyses, HIV-related symptoms were strongly associated with physical and mental health, whereas race, sex, health insurance status, disease stage, and CD4 count were at most weakly associated with physical and mental health. CONCLUSIONS: There is substantial morbidity associated with HIV disease in adults. The variability in health-related quality of life according to disease progression is relevant for health policy and allocation of resources, and merits the attention of clinicians who treat patients with HIV disease.

Reliability and validity of self-report CD4 counts-in persons hospitalized with HIV disease.

Studies of health care outcomes and clinical decision making for people with HIV disease depend on CD4 cell count data to accurately assess the stage of disease. The possibility of obtaining reliable and valid data from self-reported CD4 counts is an unexplored source of potentially important, cost-effective information for these purposes. We examined the extent of agreement of self-reported CD4 counts with medical record CD4 among 120 patients (95% male, 69% white, 5% injection drug users) hospitalized with HIV-related illness at seven Los Angeles area hospitals. Average record and report CD4 counts did not differ significantly, and record and report CD4 counts were highly correlated (product moment correlation of 0.84, intraclass correlation of 0.82). Agreement between self-reports and medical records varied by CD4 level: at higher levels of CD4, the differences between self-reports and medical records tended to be larger, with self-reports yielding upwardly biased estimates compared to the medical records. These findings suggest that self-report CD4 data may provide clinically adequate estimates of true CD4 counts. The study needs to be replicated in other populations, notably those with larger numbers of subjects who are female, of minority ethnicity, or injection drug users.

Use of dental care by HIV-infected medical patients.

Although increasing attention has been paid to the use of dental care by HIV patients, the existing studies do not use probability samples, and no accurate population estimates of use can be made from this work. The intent of the present study was to establish accurate population estimates of the use of dental services by patients under medical care. The study, part of the HIV Cost and Services Utilization Study (HCSUS), created a representative national probability sample, the first of its kind, of HIV-infected adults in medical care. Both bivariate and logistic regressions were conducted, with use of dental care in the preceding 6 months as the dependent variable and demographic, social, behavioral, and disease characteristics as independent variables. Forty-two percent of the sample had seen a dental health professional in the preceding 6 months. The bivariate logits for use of dental care show that African-Americans, those whose exposure to HIV was caused by hemophilia or blood transfusions, persons with less education, and those who were employed were less likely to use dental care (p < 0.05). Sixty-five percent of those with a usual source of care had used dental care in the preceding 6 months. Use was greatest among those obtaining dental care from an AIDS clinic (74%) and lowest among those without a usual source of dental care (12%). We conclude that, in spite of the high rate of oral disease in persons with HIV, many do not use dental care regularly, and that use varies by patient characteristics and availability of a regular source of dental care.

Evaluating the effect of usual source of dental care on access to dental services: comparisons among diverse populations.

Data collected from the International Collaborative Study of Oral Health Outcomes USA (ICS-II) research locations were used to evaluate the relationship between having a usual source of dental care (USDC) and access to dental services. The robustness of the USDC effect after reducing simultaneity bias was tested in the population-based samples using sensitivity analysis. Logistic regression results provided evidence that USDC was the strongest and most consistent predictor of a dental visit in the past 12 months regardless of geographic location, dental care delivery system, or cultural diversity of the population. Even removing cases that had a USDC for less than 1 or 2 years, the effect remained. From a policy perspective, USDC remains critical to understanding and explaining dental are utilization patterns. The findings suggest the need for designing interventions to increase the availability of a regular provider among vulnerable populations.

Breast feeding promotion in an urban and a rural Jamaican hospital.

Breast feeding promotion has been a national priority in Jamaica since the mid-1970s. Despite this effort, breast feeding rates have continued to decline there, especially in urban areas. Studies of the role of health professionals in promoting breast feeding have shown that specific practices encourage breast feeding, while others discourage it. In the context of declining breast feeding in a nation committed to promoting it, the goal of this study was to explore the relationship between specific health professional practices, mothers' breast feeding, and mothers' knowledge of breast feeding in rural and urban Jamaica. To accomplish this goal, a structured interview was administered to 113 mothers of infants age 0-6 months at one urban (n = 62) and one rural (n = 51) hospital, regarding (1) physician and nurse practices known to affect breast feeding, (2) mothers' own breast feeding practices, and (3) mothers knowledge of breast feeding issues. Physician and nurse visits were also directly observed to confirm responses given to the questionnaire and to assess the resources devoted to teaching mothers about breast feeding. While extensive lectures, posters and pamphlets promoting breast feeding were provided for mothers at the urban hospital, far fewer educational resources were available for mothers at the rural hospital. Overall, however, health professional practices discouraging breast feeding were observed more frequently at the urban hospital than at the rural hospital, whereas practices promoting breast feeding were more common at the rural hospital. At the rural hospital, mothers' breast feeding practices more closely approximated recommendations than at the urban hospital.(ABSTRACT TRUNCATED AT 250 WORDS)

Do black and white adults use the same sources of information about AIDS prevention?

Although AIDS prevention campaigns need to target population segments that are at highest risk to be effective, little is known about how various sources of AIDS information vary by race, education, and age. To determine the most common communication channels for AIDS information reported by Blacks and Whites, the authors interviewed 1,769 adults in Baltimore, Maryland, to obtain data on nine common sources of information about AIDS and analyzed their reports by race, age, and education. Television and newspapers were the most common sources but varied little across groups. National and local public health agencies, as well as medical doctors and dentists, were more commonly reported by Blacks than by Whites. Religious organizations were much more commonly reported by Blacks than by Whites. Public health organizations working collaboratively with religious organizations and health care providers might be more effective in developing AIDS prevention strategies than has been considered previously.

Health status measurement performance and health status differences by age, ethnicity, and gender: assessment in the medical outcomes study.

The comparative measurement performance of self-reported health status instruments for African American and Hispanic elderly has rarely been studied, despite evidence of their poor health status. This study examined psychometric performance and health status differences by age, ethnicity, and gender among 10,569 ethnically diverse patients who completed the Short-Form General Health Survey in the Medical Outcomes Study (MOS). Hispanics and African Americans tended to have slightly lower measurement performance than other ethnic groups. Compared with whites, health status scores for African American and Hispanic women were slightly but significantly lower. The small differences in health status by ethnicity may be due to the MOS sampling strategy, which excluded low-socioeconomic status minorities with poor education and no regular medical care. The psychometric performance of MOS health status measures should be examined in studies of ethnically diverse, community-dwelling, elderly populations who have poor access to care, poor education, and/or low socioeconomic status.

Ethnic and racial differences in long-term survival from hospitalization for HIV infection.

This prospective cohort study compares 200 hospitalized, HIV-infected patients (Hispanic, African American, and white) from May 1992 to October 1998 to assess mortality (versus survival) over 75 months of follow-up. The relative risk of six-year mortality for each ethnic group is compared using Cox proportional hazards models after controlling for sociodemographic and clinical characteristics, access to general medical care, and HIV-specific treatment. The median survival of Hispanics (15.5 months) was significantly (p < 0.05) shorter than that of whites (23.8); survival for African Americans (35.1) did not differ from whites. In multivariate analysis, the adjusted relative risk of six-year mortality for Hispanics compared with whites was 2.14 (95 percent confidence interval = 1.26-3.66). The poor outcomes of Hispanics was not explained by access to general care or by HIV-specific treatment.

A comparison of unmet needs for dental and medical care among persons with HIV infection receiving care in the United States.

OBJECTIVE: Oral health conditions associated with HIV disease are frequently more severe than those of the general population, making access to both dental and medical care important. Using the domains specified in the Behavioral Model of Health Services Use, this paper examines the correlates of unmet needs for dental and medical care in a nationally representative sample of patients with HIV. METHODS: This investigation is a cross-sectional study using baseline data from the HIV Cost and Services Utilization Study (HCSUS), the first nationally representative study of persons in care for HIV. Using probability-based techniques, 4,042 people were randomly selected in January 1996, and 2,864 (71%) completed a structured interview that included questions on unmet needs for dental and medical care. Regression analysis was used to identify variables associated with having unmet needs for dental care only, medical care only, and both medical and dental care. RESULTS: Of the estimated 230,900 people in treatment for HIV in the United States, approximately 58,000 had unmet medical or dental needs based on self-reported data. Unmet dental needs were more than twice as prevalent as unmet medical needs (32,900 vs 14,300), and 11,600 people were estimated to have both types of unmet needs. Multinomial logit regression showed that persons with low income had increased odds of reporting unmet needs for both dental and medical care. The uninsured and those insured by Medicaid without dental benefits had more than three times the odds of having unmet needs for both types of care than did the privately insured. CONCLUSIONS: To serve both the dental and medical needs of diverse populations affected by HIV disease, greater coordination of services is needed. In addition, state insurance programs for people with HIV should consider the feasibility of expanding their benefit structure to include dental care benefits.

Use of conjoint analysis to assess HIV vaccine acceptability: feasibility of an innovation in the assessment of consumer health-care preferences.

Engaging consumers in prospectively shaping strategies for dissemination of health-care innovations may help to ensure acceptability. We examined the feasibility of using conjoint analysis to assess future HIV vaccine acceptability among three diverse communities: a multiethnic sample in Los Angeles, CA, USA (n = 143); a Thai resident sample in Los Angeles (three groups; n = 27) and an Aboriginal peoples sample in Toronto (n = 13). Efficacy had the greatest impact on acceptability for all three groups, followed by cross-clade protection, side-effects and duration of protection in the Los Angeles sample; side-effects and duration of protection in the Thai-Los Angeles sample; and number of doses and duration of protection in the Aboriginal peoples-Toronto sample. Conjoint analysis provided insights into universal and population-specific preferences among diverse end users of future HIV vaccines, with implications for evidence-informed targeting of dissemination efforts to optimize vaccine uptake.

Acceptability of pre-exposure prophylaxis as an HIV prevention strategy: barriers and facilitators to pre-exposure prophylaxis uptake among at-risk Peruvian populations.

This study examined pre-exposure prophylaxis (PrEP) acceptability among female sex workers, male-to-female transgendered persons and men who have sex with men in Lima, Peru. Focus groups explored social issues associated with PrEP acceptability and conjoint analysis assessed preferences among eight hypothetical PrEP scenarios with varying attribute profiles and their relative impact on acceptability. Conjoint analysis revealed that PrEP acceptability ranged from 19.8 to 82.5 out of a possible score of 100 across the eight hypothetical PrEP scenarios. Out-of-pocket cost had the greatest impact on PrEP acceptability (25.2, P < 0.001), followed by efficacy (21.4, P < 0.001) and potential side-effects (14.7, P < 0.001). Focus group data supported these findings, and also revealed that potential sexual risk disinhibition, stigma and discrimination associated with PrEP use, and mistrust of health-care professionals were also concerns. These issues will require careful attention when planning for PrEP roll-out.

Associations of clinical parameters with health-related quality of life in hospitalized persons with HIV disease.

To assess the associations of clinical parameters with health-related quality of life (HRQOL) in hospitalized patients with HIV a cross-sectional survey and medical record review was conducted. Medical record reviews for clinical data were completed for 217 hospitalized, HIV-positive individuals. HRQOL was measured in face-to-face interviews using a 44-item survey instrument that yielded 11 multi-item scales. Internal consistency reliability was adequate for the 11 HRQOL scales. HRQOL scores were significantly lower for our hospitalized sample compared to a previous sample of ambulatory clinical trial patients. As expected, individuals reporting more days spent in bed due to illness had significantly lower HRQOL than those reporting fewer days spent in bed due to illness. In addition, those with greater severity of illness had significantly lower HRQOL than persons with less severe illness. In multivariate models, severity of illness and the number of disability days due to illness were significantly associated with worse HRQOL, controlling for CD4 count, symptoms and other patient characteristics. Severity of illness and disability days due to illness are associated with poorer functioning and well-being in hospitalized persons with HIV disease. HRQOL measures are useful tools for assessing the effects of HIV disease on the day-to-day life of individuals with HIV disease.

Are Latinos less satisfied with communication by health care providers?

OBJECTIVE: To examine associations of patient ratings of communication by health care providers with patient language (English vs Spanish) and ethnicity (Latino vs white). METHODS: A random sample of patients receiving medical care from a physician group association concentrated on the West Coast was studied. A total of 7,093 English and Spanish language questionnaires were returned for an overall response rate of 59%. Five questions asking patients to rate communication by their health care providers were examined in this study. All five questions were administered with a 7-point response scale. MAIN RESULTS: We estimated the associations of satisfaction ratings with language (English vs Spanish) and ethnicity (white vs Latino) using ordinal logistic models, controlling for age and gender. Latinos responding in Spanish (Latino/Spanish) were significantly more dissatisfied compared with Latinos responding in English (Latino/English) and non-Latino whites responding in English (white) when asked about: (1) the medical staff listened to what they say (29% vs 17% vs 13% rated this "very poor," "poor," or "fair"; p <.01); (2) answers to their questions (27% vs 16% vs 12%; p <.01); (3) explanations about prescribed medications (22% vs 19% vs 14%; p <.01); (4) explanations about medical procedures and test results (36% vs 21% vs 17%; p <.01); and (5) reassurance and support from their doctors and the office staff (37% vs 23% vs 18%; p <.01). CONCLUSION: This study documents that Latino/Spanish respondents are significantly more dissatisfied with provider communication than Latino/English and white respondents. These results suggest Spanish-speaking Latinos may be at increased risk of lower quality of care and poor health outcomes. Efforts to improve the quality of communication with Spanish-speaking Latino patients in outpatient health care settings are needed.

Access to medical care and health-related quality of life for low-income persons with symptomatic human immunodeficiency virus.

Despite growing interest in the accessibility of medical care and health-related quality of life for persons infected with human immunodeficiency virus, an association between these variables has not been documented. The authors conducted a cross-sectional study of access to care and its association with health-related quality of life among 205 persons of low income infected with the human immunodeficiency virus with constitutional symptoms and/or diarrhea at one public and one Veterans Administration hospital, using a 9-item measure of perceived access and a 55-item health-related quality of life instrument. Problems with access were widespread: 55% traveled for longer than 30 minutes to their usual source of care (compared with 9% to 12% of general populations in national surveys), 49% had problems meeting costs of care, and 48% had problems with clinic hours (compared with 23% in national surveys). In multivariate analyses, uninsured patients reported worse access than patients with Medicaid or Veterans Administration insurance, particularly for meeting the cost of care (P < 0.01). Adjusted health-related quality of life scores in this sample were far lower (by about 1 SD) than those of subjects in a large national acquired immune deficiency syndrome clinical trial. For 8 of 11 health-related quality of life subscales, worse perceived access was significantly (P < 0.05) associated with poorer health-related quality of life, even after controlling for T-4 lymphocyte count, symptoms and other factors. Access and health-related quality of life measures similar to those used in this study may prove useful in future evaluations of medical care systems serving poor, clinically ill populations infected with human immunodeficiency virus.

Use of relevancy ratings by target respondents to develop health-related quality of life measures: an example with African-American elderly.

BACKGROUND: Health-related quality of life (HRQOL) instruments assess functioning and well-being. Generic HRQOL measures are intended to be relevant to everyone whereas population-targeted measures are designed to be relevant to a particular population. METHODS: We asked 99 African-American elderly (mean age 72, 33% female, 47% less than high school education) to rate the relevancy of 33 HRQOL items drawn largely from existing instruments. We assessed the reliability of the relevancy ratings across respondents, rank-ordered the items by relevancy, and tested the significance of difference in relevancy ratings for each item compared to the average of all other items. We also examined the associations of the relevancy ratings with sociodemographic and clinical characteristics. RESULTS: The relevancy ratings were reliable (intraclass correlation = 0.71) and relevancy was generally distinct from HRQOL and demographic characteristics. Items assessing spirituality and weight-related health status were rated as significantly more relevant than other types of items. Generic HRQOL items were not rated as highly relevant. CONCLUSIONS: HRQOL measures assessing spirituality and weight-related concepts are important for future studies of HRQOL in African-American elderly. The method of identifying these concepts used in this study should be valuable in developing new measures targeted to other sociodemographically or clinically defined subgroups.

Is access to medical care associated with receipt of HIV testing and counselling?

Lack of timely HIV testing leads to missed prevention opportunities and poor prevention counselling may be related to further disease spread. We examined the association of self-reported access to medical care with receiving HIV testing and preventive counselling services among a sample of patients with HIV disease prior to hospitalization. We conducted a cross-sectional interview of 217 Los Angeles patients hospitalized with HIV-related illness between 1992 and 1993 and abstracted clinical data from the medical record. Eighty-four per cent of patients received HIV testing prior to hospitalization, but only 33% received preventive counselling services. Only 48% of all patients rated outpatient medical care as somewhat or very easy to obtain. Controlling for severity of illness, better access to outpatient medical care (OR = 1.48; 95% CI = 1.02-2.15), having a regular source of care (OR = 3.40; 95% CI = 1.29-8.97) and non-homosexual mode of HIV transmission (OR = 0.31; 0.12-0.83) were associated with receiving HIV testing services prior to hospitalization. Having a regular source of care (OR = 3.55; 95% CI = 1.37-9.22), being VA (Veterans' Administration) insured (OR = 6.16; 1.46-26.05), older age (OR = 0.95; 95% CI = 0.90-0.99) and having a CD4 count between 101-200 (OR = 0.19; 95% CI = 0.06-0.63) were associated with receiving HIV counselling. Limited self-reported access to medical care is associated with fewer patients receiving HIV testing and counselling. Improving timeliness of HIV testing may require removing the barriers to medical care.

Comparison of health-related quality of life in clinical trial and nonclinical trial human immunodeficiency virus-infected cohorts.

Clinical trials randomly assign treatments and select participants to maximize internal validity, but such selection threatens generalizability by excluding important groups with the diseases under study. Particularly in human immunodeficiency virus (HIV) disease, the results of clinical trials are applied broadly to populations, despite limited representation by minorities and disadvantaged groups. Health-related quality of life (HRQOL), which is increasingly recognized as an important outcome in these studies, may be sensitive to differences that affect generalization of trial results to target populations. This study compared HRQOL in two HIV-infected cohorts: 1) multicenter AIDS Clinical Group Trials in which most subjects are white, privately insured, and high-income (n = 1,907); and 2) a study of ethnically diverse, low-income patients recruited from public clinics (n = 205). Both studies included 30 HRQOL items developed in the Medical Outcomes Study (MOS) and items on symptoms, medications, and demographic characteristics. HRQOL scores were significantly lower in the nontrial sample (P < 0.001) by about one standard deviation, even after direct adjustment for clinical and demographic characteristics, and also after comparison of the nontrial sample with the most symptomatic in the trial sample. The relationships of characteristics with HRQOL differed between nontrial and trial samples, suggesting problems generalizing results from HIV clinical trials to important target populations. HRQOL measures such as those from the MOS can be useful in detecting differences that affect generalization.