Perceptions and acceptability of the controlled human malaria infection (CHMI) model in The Gambia: a qualitative study.

Controlled human malaria infection (CHMI) studies, i.e. the deliberate infection of healthy volunteers with malaria parasites to study immune response and/or test drug or vaccine efficacy, are increasingly being conducted in malaria endemic countries, including in sub-Saharan Africa. However, there have been few studies on the perceptions and acceptability of CHMI by the local communities. This qualitative study assessed the perception and acceptability of such studies in The Gambia following the first CHMI study conducted in the country in March-May 2018. Data were collected through non-participant observation, in-depth interviews and focus group discussions and analyzed using NVivo 12 software with an inductive-deductive approach. Sixty-seven participants were involved, including volunteers enrolled in the CHMI, community stakeholders and members of the Gambian Ethics Committee. Respondents expressed a positive view about CHMI. Key motivating factors for participation were the financial compensation, comprehensive health checks, and willingness to support malaria research. Risks associated with participation were considered low. Concerns raised included the frequency of bleeding and the blood volume collected.

Making global health 'work': Frontline workers' labour in research and interventions.

This Special Issue of Global Public Health draws on the concept of 'body work' among those employed to support operationalising, researching, and implementing global health while in direct contact with the bodies of others. This collection brings into sharp focus the specific forms of labour of those occupying positions as frontline workers - those who make global health work. Making Global Health Work includes authors from diverse backgrounds, disciplines, and geographies. Through compelling ethnographies, qualitative interviews, and focus group discussions, they explore 'body work' globally, including: Afghanistan, Bangladesh, Ethiopia, India, Indonesia, Kenya, Malawi, Myanmar, Nigeria, Nepal, Pakistan, Sierra Leone, South Sudan, Tanzania, Thailand, The Democratic Republic of the Congo (DRC), The Gambia, Vietnam, and Zimbabwe. These papers demonstrate that proximity to, and work on, the bodies of others engenders specific forms of (physical, emotional, mental, social, ethical, and political) labour, which occur not only in emergencies and pandemics, but also throughout the quotidian practice of global health. Making Global Health Work provides insights into the provision of maternal healthcare, treatment of multidrug resistant tuberculosis, rapid HIV testing programmes, sleeping sickness and polio eradication campaigns, mass drug administration clinical trials, epidemic preparedness and response, and the management and care of dead bodies. These papers argue for greater attention by global health actors on frontline workers in management of the complexities involved in making global health work.

Doing 'reciprocity work': The role of fieldworkers in a mass drug administration trial in the Gambia.

In their roles as nurses, data collectors, or other, fieldworkers undertake myriad tasks working intimately with and on the bodies of others - a type of work called 'body work'. This work further includes the micro-political relations shaping these interactions, and studies have shown the importance of these relationships in the success of clinical trials, particularly in the Gambia. This study seeks to expand the concept of body work to understand the roles and interactions of fieldworkers within the trial community, and the effect on a mass drug administration (MDA) clinical trial. We conducted a mixed-methods social science study alongside the MDA in 2018-2019, including in-depth interviews, focus group discussions, and semi-structured observations with the population involved (and not) in the MDA, as well as the MRC fieldworkers. We found that fieldworkers participated in what we call 'reciprocity work'. Through their regular tasks and interactions, they necessarily showed respect and established trust in a way that formed and contributed to reciprocal relationships, the results of which impacted the trial and individuals' autonomy in the decision-making process. Understanding the role of fieldworkers and their reciprocity work is a vital component in comprehending how research ethics are made and conducted in global health research.

The role of social cohesion in the implementation and coverage of a mass drug administration trial for malaria control in the Gambia: An in-depth comparison of two intervention villages.

Mass drug administration (MDA), used increasingly in malaria eradication efforts, involves administering medication to an entire target population regardless of individual-level disease status. This strategy requires high levels of coverage and compliance. Previous studies have assessed individual and structural factors affecting MDA coverage, but there is a need to better understand the influence and expressions of community dynamics and social structures, such as social cohesion. We conducted a social science study concurrent to an MDA clinical trial for malaria control in The Gambia; ethnographic research was conducted prior to, throughout, and between MDA implementation July-November 2018, January-March 2019, and July-November 2019. We assessed how social cohesion, as expressed by the trial population, affects trial coverage through an in-depth ethnographic analysis of two trial villages, using observations, interviews, and focus group discussions with community members who took the trial medication and those who did not. We found that the villages had unique expressions of social cohesion. This was reflected through community participation in the trial implementation and may have affected coverage and compliance. The village with low coverage expressed a form of social cohesion where members followed advice to participate through a hierarchal system but did not actively participate in the MDA or its implementation. The village with high coverage expressed social cohesion as more participatory: individuals took the directive to participate but contextualized the trial implementation to their needs and wants. We analyze these different expressions of social cohesion and the important differences they make for the coverage and compliance levels reached in the two different villages.

Costs and barriers faced by households seeking malaria treatment in the Upper River Region, The Gambia.

BACKGROUND: Malaria transmission in The Gambia decreased substantially over the last 20 years thanks to the scale-up of control interventions. However, malaria prevalence is still relatively high in eastern Gambia and represents both a health and a financial burden for households. This study aims to quantify the out-of-pocket costs and productivity losses of seeking malaria treatment at household level. METHODS: A household survey was carried out through in-person interviews. Respondents were asked about malaria prevention methods, their treatment-seeking behaviour, and any costs incurred for transport, services, food, and/or overnight stays. A bottom-up costing approach was used to calculate the unit cost of treatment and a tobit regression approach to investigate cost drivers. RESULTS: The survey included 864 respondents, mainly subsistence farmers. Most respondents (87%) considered malaria to be a problem affecting their ability to perform their regular duties. Respondents preferred going to a health facility for treatment. The primary reason for not going was related to costs; 70% of respondents incurred costs for seeking health care, with a median of £3.62 (IQR: £1.73 to £6.10). The primary driver of cost was living in one of the villages that are off the main road and/or far from health facilities. 66% reported productivity loss of 5 working days on average during a malaria episode of them or their child. CONCLUSIONS: Although malaria prevalence is decreasing and treatment is provided free of charge, households seeking treatment are confronted with out-of-pocket expenditures and lost working days; particularly in remote villages.

From informed consent to adherence: factors influencing involvement in mass drug administration with ivermectin for malaria elimination in The Gambia.

BACKGROUND: The World Health Organization (WHO) recommends consideration of mass drug administration (MDA) for malaria control in low-endemic settings approaching elimination. However, MDA remains a controversial strategy, as multiple individual, social, and operational factors have shown to affect its acceptability at local levels. This is further complicated by inconsistent definitions of key indicators derived from individual and community involvement-coverage, adherence, and compliance-that cast doubts about the actual and potential epidemiological impact of MDA on disease control and elimination. This study aimed to identify limitations and enabling factors impacting involvement at different stages of a large cluster-randomized trial assessing the effect of combining dihydroartemisinin-piperaquine (DP) and ivermectin (IVM) in malaria transmission in The Gambia. METHODS: This social science study used a mixed-methods approach. Qualitative data were collected in intervention and control villages through ethnographic methods, including in-depth interviews (IDIs), focus group discussions (FGDs), and participant observation conducted with trial participants and decliners, community leaders, and field staff. A cross-sectional survey was conducted in the intervention villages after the first year of MDA. Both strands of the study explored malaria knowledge and opinions, social dynamics influencing decision-making, as well as perceived risks, burdens, and benefits associated with this MDA. RESULTS: 157 IDIs and 11 FGDs were conducted, and 864 respondents were included in the survey. Barriers and enabling factors to involvement were differentially influential at the various stages of the MDA. Issues of social influence, concerns regarding secondary effects of the medication, costs associated with malaria, and acceptability of the implementing organization, among other factors, differently affected the decision-making processes throughout the trial. Rather than a linear trajectory, involvement in this MDA trial was subjected to multiple revaluations from enrolment and consent to medicine intake and adherence to treatment. CONCLUSIONS: This study went beyond the individual factors often associated with coverage and adherence, and found that nuanced social dynamics greatly influence the decision-making process at all phases of the trial. These issues need to be consider for MDA implementation strategies and inform discussions about more accurate ways of reporting on critical effectiveness indicators.

Mass Drug Administration With High-Dose Ivermectin and Dihydroartemisinin-Piperaquine for Malaria Elimination in an Area of Low Transmission With High Coverage of Malaria Control Interventions: Protocol for the MASSIV Cluster Randomized Clinical Trial.

BACKGROUND: With a decline in malaria burden, innovative interventions and tools are required to reduce malaria transmission further. Mass drug administration (MDA) of artemisinin-based combination therapy (ACT) has been identified as a potential tool to further reduce malaria transmission, where coverage of vector control interventions is already high. However, the impact is limited in time. Combining an ACT with an endectocide treatment that is able to reduce vector survival, such as ivermectin (IVM), could increase the impact of MDA and offer a new tool to reduce malaria transmission. OBJECTIVE: The study objective is to evaluate the impact of MDA with IVM plus dihydroartemisinin-piperaquine (DP) on malaria transmission in an area with high coverage of malaria control interventions. METHODS: The study is a cluster randomized trial in the Upper River Region of The Gambia and included 32 villages (16 control and 16 intervention). A buffer zone of ~2 km was created around all intervention clusters. MDA with IVM plus DP was implemented in all intervention villages and the buffer zones; control villages received standard malaria interventions according to the Gambian National Malaria Control Program plans. RESULTS: The MDA campaigns were carried out from August to October 2018 for the first year and from July to September 2019 for the second year. Statistical analysis will commence once the database is completed, cleaned, and locked. CONCLUSIONS: This is the first cluster randomized clinical trial of MDA with IVM plus DP. The results will provide evidence on the impact of MDA with IVM plus DP on malaria transmission. TRIAL REGISTRATION: ClinicalTrials.gov NCT03576313; https://clinicaltrials.gov/ct2/show/NCT03576313. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/20904.

Factors Associated with Loss to Follow-up among Cervical Cancer Patients in Rwanda.

Background: Cervical cancer is among the most common cancers affecting women globally. Where treatment is available in low- and middle-income countries, many women become lost to follow-up (LTFU) at various points of care. Objective: This study assessed predictors of LTFU among cervical cancer patients in rural Rwanda. Methods: We conducted a retrospective study of cervical cancer patients enrolled at Butaro Cancer Center of Excellence (BCCOE) between 2012 and 2017 who were either alive and in care or LTFU at 12 months after enrollment. Patients are considered early LTFU if they did not return to clinic after the first visit and late LTFU if they did not return to clinic after the second visit. We conducted two multivariable logistic regressions to determine predictors of early and late LTFU. Findings: Of 652 patients in the program, 312 women met inclusion criteria, of whom 47 (15.1%) were early LTFU, 78 (25.0%) were late LTFU and 187 (59.9%) were alive and in care. In adjusted analyses, patients with no documented disease stage at presentation were more likely to be early LTFU vs. patients with stage 1 and 2 when controlling for other factors (aOR: 14.93, 95% CI 6.12-36.43). Patients who travel long distances (aOR: 2.25, 95% CI 1.11, 4.53), with palliative care as type of treatment received (aOR: 6.65, CI 2.28, 19.40) and patients with missing treatment (aOR: 7.99, CI 3.56, 17.97) were more likely to be late LTFU when controlling for other factors. Patients with ECOG status of 2 and higher were less likely to be late LTFU (aOR: 0.26, 95% CI 0.08, 0.85). Conclusion: Different factors were associated with early and later LTFU. Enhanced patient education, mechanisms to facilitate diagnosis at early stages of disease, and strategies that improve patient tracking and follow-up may reduce LTFU and improve patient retention.

Patient Characteristics, Early Outcomes, and Implementation Lessons of Cervical Cancer Treatment Services in Rural Rwanda.

PURPOSE: Low- and middle-income countries account for 86% of all cervical cancer cases and 88% of cervical cancer mortality globally. Successful management of cervical cancer requires resources that are scarce in sub-Saharan Africa, especially in rural settings. Here, we describe the early clinical outcomes and implementation lessons learned from the Rwanda Ministry of Health's first national cancer referral center, the Butaro Cancer Center of Excellence (BCCOE). We hypothesize that those patients presenting at earlier stage and receiving treatment will have higher rates of being alive. METHODS: The implementation of cervical cancer services included developing partnerships, clinical protocols, pathology services, and tools for monitoring and evaluation. We conducted a retrospective study of patients with cervical cancer who presented at BCCOE between July 1, 2012, and June 30, 2015. Data were collected from the electronic medical record system and by manually reviewing medical records. Descriptive, bivariable and multivariable statistical analyses were conducted to describe patient demographics, disease profiles, treatment, and clinical outcomes. RESULTS: In all, 373 patients met the study inclusion criteria. The median age was 53 years (interquartile rage, 45 to 60 years), and 98% were residents of Rwanda. Eighty-nine percent of patients had a documented disease stage: 3% were stage I, 48% were stage II, 29% were stage III, and 8% were stage IV at presentation. Fifty percent of patients were planned to be treated with a curative intent, and 54% were referred to chemoradiotherapy in Uganda. Forty percent of patients who received chemoradiotherapy were in remission. Overall, 25% were lost to follow-up. CONCLUSION: BCCOE illustrates the feasibility and challenges of implementing effective cervical cancer treatment services in a rural setting in a low-income country.

Impact of delayed care on surgical management of patients with gastric cancer in a low-resource setting.

BACKGROUND: Gastric cancer is the fifth most common cancer in Eastern Africa. Diagnostic delays in low-resource countries result in advanced disease presentation. We describe perioperative management of gastric cancer in Rwanda. METHODS: A retrospective review of records at three hospitals was performed to identify gastric adenocarcinoma cases from January 2012 to June 2016. Multiple perioperative and tumor-related variables were collected. Descriptive and bivariate analyses were performed. RESULTS: The final analysis included 229 patients with gastric cancer. Median age was 58 years (interquartile range [IQR] 49-65) and 49.6% were female (n = 114). Patients reported symptoms (ie, weight loss, epigastric pain) for a median time of 12 months (IQR 7.5-24). On presentation, 18.8% ( n = 43) had gastric outlet obstruction; 13.5% ( n = 31) had a palpable mass. Fifty-one percent ( n = 117) underwent an operation; of these, 74% ( n = 86) received gastrojejunostomy or were inoperable; and 29% ( n = 34) underwent curative resection. Palliative care referrals were made for 9% ( n = 20). Pathology reports were available for 190 patients (83.0%). Only 11.3% ( n = 26) had Helicobacter pylori ( H. pylori) testing of which 65.4% tested positive ( n = 17). CONCLUSIONS: A majority of patients presented with advanced disease. Very few patients had a curative resection. Significant advances in diagnosis and treatment are needed to improve the care of gastric cancer patients in Rwanda.

International Radiotherapy Referrals From Rural Rwanda: Implementation Processes and Early Clinical Outcomes.

PURPOSE: Low- and middle-income countries disproportionately comprise 65% of cancer deaths. Cancer care delivery in resource-limited settings, especially low-income countries in sub-Saharan Africa, is exceedingly complex, requiring multiple modalities of diagnosis and treatment. Given the vast human, technical, and financial resources required, access to radiotherapy remains limited in sub-Saharan Africa. Through 2017, Rwanda has not had in-country radiotherapy services. The aim of this study was to describe the implementation and early outcomes of the radiotherapy referral program at the Butaro Cancer Centre of Excellence and to identify both successful pathways and barriers to care. METHODS: Butaro District Hospital is located in a rural area of the Northern Province and is home to the Butaro Cancer Centre of Excellence. We performed a retrospective study from routinely collected data of all patients with a diagnosis of cervical, head and neck, or rectal cancer between July 2012 and June 2015. RESULTS: Between 2012 and 2015, 580 patients were identified with these diagnoses and were potential candidates for radiation. Two hundred eight (36%) were referred for radiotherapy treatment in Uganda. Of those referred, 160 (77%) had cervical cancer, 31 (15%) had head and neck cancer, and 17 (8%) had rectal cancer. At the time of data collection, 101 radiotherapy patients (49%) were alive and had completed treatment with no evidence of recurrence, 11 (5%) were alive and continuing treatment, and 12 (6%) were alive and had completed treatment with evidence of recurrence. CONCLUSION: This study demonstrates the feasibility of a rural cancer facility to successfully conduct out-of-country radiotherapy referrals with promising early outcomes. The results of this study also highlight the many challenges and lessons learned in providing comprehensive cancer care in resource-limited settings.

Decreasing Histology Turnaround Time Through Stepwise Innovation and Capacity Building in Rwanda.

PURPOSE: Minimal turnaround time for pathology results is crucial for highest-quality patient care in all settings, especially in low- and middle-income countries, where rural populations may have limited access to health care. METHODS: We retrospectively determined the turnaround times (TATs) for anatomic pathology specimens, comparing three different modes of operation that occurred throughout the development and implementation of our pathology laboratory at the Butaro Cancer Center of Excellence in Rwanda. Before opening this laboratory, TAT was measured in months because of inconsistent laboratory operations and a paucity of in-country pathologists. RESULTS: We analyzed 2,514 individual patient samples across the three modes of study. Diagnostic mode 1 (samples sent out of the country for analysis) had the highest median TAT, with an overall time of 30 days (interquartile range [IQR], 22 to 43 days). For diagnostic mode 2 (static image telepathology), the median TAT was 14 days (IQR, 7 to 27 days), and for diagnostic mode 3 (onsite expert diagnosis), it was 5 days (IQR, 2 to 9 days). CONCLUSION: Our results demonstrate that telepathology is a significant improvement over external expert review and can greatly assist sites in improving their TATs until pathologists are on site.

Outcomes of Low-Intensity Treatment of Acute Lymphoblastic Leukemia at Butaro Cancer Center of Excellence in Rwanda.

PURPOSE: Children with acute lymphoblastic leukemia (ALL) in low-income countries have disproportionately lower cure rates than those in high-income countries. At Butaro Cancer Center of Excellence (BCCOE), physicians treated patients with ALL with the first arm of the Hunger Protocol, a graduated-intensity method tailored for resource-limited settings. This article provides the first published outcomes, to our knowledge, of patients with ALL treated with this protocol. METHODS: This is a retrospective descriptive study of patients with ALL enrolled at BCCOE from July 1, 2012 to June 30, 2014; data were collected through December 31, 2015. Descriptive statistics were used to calculate patient demographics, disease characteristics, and outcomes; event-free survival was assessed at 2 years using the Kaplan-Meier method. RESULTS: Forty-two consecutive patients with ALL were included. At the end of the study period, 19% (eight) were alive without evidence of relapse: three completed treatment and five were continuing treatment. Among the remaining patients, 71% (30) had died and 10% (four) were lost to follow-up. A total of 83% (25) of the deaths were disease related, 3% (one) treatment-related, and 13% (four) unclear. Event-free survival was 22% (95% CI, 11% to 36%), considering lost to follow-up as an event, and 26% (95% CI, 13% to 41%) if lost to follow-up is censored. CONCLUSION: As expected, relapse was the major cause of failure with this low-intensity regimen. However, toxicity was acceptably low, and BCCOE has decided to advance to intensity level 2. These results reflect the necessity of a data-driven approach and a continual improvement process to care for complex patients in resource-constrained settings.

Patient and Surrogate Views of Community Consultation for Emergency Research.

OBJECTIVE: Pretrial community consultation (CC) is required for emergency research conducted under an exception from informed consent (EFIC) in the United States. CC remains controversial and challenging, and minimal data exist regarding the views of individuals enrolled in EFIC trials on this process. It is important to know whether participants perceive CC to be meaningful and, if so, whom they believe should be consulted. METHODS: We conducted a secondary analysis of data from two studies interviewing patients and surrogates of two recent EFIC trials (PEER-RAMPART and PEER-ProTECT). These interviews included similar open- and closed-ended questions regarding participants' views of the importance of CC, the rationale for their responses, and their views regarding which populations should be included in consultation efforts. A template analytic strategy was used for qualitative analysis of textual data, and descriptive statistics were tabulated to characterize demographic data and instances of major themes. RESULTS: Ninety percent of participants perceived CC to be valuable. Participants' reasons for finding CC valuable clustered in two categories: 1) as a method of informing the public about the trial to be conducted and 2) as a way of obtaining input and feedback from the community. Participants cited the medical community (43%) and individuals with a connection to the study condition (41%) as the most important groups to involve in consultation efforts; only 5% suggested consulting the general public in the area where the research will be conducted. CONCLUSION: Participants in EFIC trials and their decision makers generally valued CC as a method of informing and seeking input from the community. Participants felt that the most appropriate groups to consult were the medical community and individuals with connections to the condition under study. Consultation efforts focused on these two groups, rather than the general public, may be more efficient and more meaningful to individuals involved in EFIC trials. These findings also reinforce the importance of the distinction between public disclosure and CC.

Patients' views of consent in clinical trials for acute myocardial infarction: impact of trial design.

OBJECTIVE: Seeking prospective informed consent is difficult in clinical trials for emergent conditions such as acute myocardial infarction (AMI). Prior data suggest that enrolment decisions of patients are often poorly informed in AMI trials but that patients prefer to be asked permission before enrolment. It is unknown whether this is true across trial designs or in comparative effectiveness research (CER) with approved treatments. METHODS: Structured interviews were conducted with 30 patients with AMI. Participants considered three scenarios: (1) a CER trial of approved antiplatelet drugs; (2) a placebo-controlled trial of a novel drug to reduce myocardial injury and (3) a CER trial of an intra-aortic balloon pump versus medication. Participants were asked their desired involvement in enrolment decisions and willingness to participate. Descriptive analysis was performed of Likert scale data, and qualitative descriptive analysis was performed of textual data. RESULTS: Across scenarios, most participants (73%-80%) preferred to be asked permission prior to trial enrolment. Reasons for involvement included wanting to be the decision maker and a desire for transparency. Willingness to enrol was affected by trial type. Fewer participants stated they would likely enrol in a CER procedural trial than in a CER trial of approved medications (p=0.012). CONCLUSIONS: These findings suggest that patients prefer prospective involvement in enrolment decisions to enrolment without consent across trial types. However, their desire to participate was affected by trial type. There is a need to develop and evaluate context-sensitive approaches to consent in AMI trials that account for both the acuity of the situation and trial characteristics.

Patients' views of consent for research enrollment during acute myocardial infarction.

OBJECTIVE: The goal of this study was to improve understanding of patients' perspectives of informed consent for clinical trial enrollment during ST-elevation myocardial infarction (STEMI). BACKGROUND: STEMI research poses challenges regarding informed consent due to time constraints, symptom severity, and potential cognitive impairment. METHODS: An interview study was embedded within a clinical trial of an ischemic post-conditioning procedure for STEMI. An interactive, structured interview guide was used. RESULTS: Twenty interviews were conducted (median 1.5 days after initial procedure). Only 11/20 (55%) initially remembered being asked to participate in a trial. Comprehension of study details and the research process was limited; for example, many patients believed their decision was primarily clinical or that the study intervention was clearly superior. Patients described limitations regarding their ability to make a decision, including pain, discomfort, lack of time, and stress. However, patients generally felt they were able to make a decision at presentation and wanted to be the primary decision maker; few supported surrogate decision-making by family or physicians. CONCLUSIONS: This study demonstrates that patients recognize barriers to understanding but value elements of consent. Further research is needed to develop approaches that meet patients' goals while recognizing the limitations inherent to this context.

Learning from experience: a systematic review of community consultation acceptance data.

STUDY OBJECTIVE: Federal regulations permitting an exception from informed consent for research in emergency settings require community consultation before study approval. Rates of acceptance of exception from informed consent in community consultation are often reported, but predictors of acceptance are not well understood, and investigators and institutional review boards struggle to interpret and use acceptance data. METHODS: We systematically reviewed empirical literature on community consultation for exception from informed consent trials in the PubMed, EMBASE, and Web of Science databases. We included peer-reviewed articles reporting acceptance data from community consultation for US exception from informed consent trials. Questions were categorized by enrollment focus (eg, personal enrollment versus more general exception from informed consent acceptance), and observed acceptance was compared across studies. We also compared potential predictors of acceptance, including demographic factors, consultation method, and target community. RESULTS: Nine studies (total n=9,036 participants) were included in the final analysis. Personal acceptance of enrollment in the proposed exception from informed consent study ranged from 45% to 93% and clustered in the range of 64% to 80%. Acceptance of the exception from informed consent mechanism in general (without reference to personal inclusion) was lower (35% to 84%) than personal acceptance. The effect of demographic characteristics on acceptance was inconsistent, and meeting-based consultation methods were associated with greater acceptance than survey-based methods. Finally, acceptance rates varied substantially according to the phrasing of the question. CONCLUSION: Personal acceptance clustered between 64% and 80%. This range may be informative for institutional review boards and investigators evaluating community consultation results. However, numerous factors affect acceptance, and there is a need for considerable caution against overreliance on acceptance data.