Uncovering the Risks of Anticancer Therapy Through Incident Report Analysis Using a Newly Developed Medical Oncology Incident Taxonomy.

BACKGROUND: Incident reporting systems were developed to identify possible and actual harm in healthcare facilities. They have the potential to capture important safety trends and to enable improvements that can mitigate the risk of future patient harm and suffering. We recently developed and validated a taxonomy specific for medical oncology designed to enhance the identification, tracking, and trending of incidents that may lead to patient harm. The current project was designed to test the ability of such a taxonomy to be applied across different organizations delivering medical oncology care and to identify specific risks that could result in future harm. METHODS: We analyzed 309 randomly selected medical oncology-related incident reports from 3 different cancer centers that had been posted between January 2019 and December 2020. Each report was assigned up to 2 incident categories. We used a 2-step process to reconcile reviewer discrepancies. In a secondary analysis, each of the incidents was reviewed and recoded to identify events which may result in major or catastrophic harm. RESULTS: Three hundred four incidents met criteria for inclusion. Three hundred incidents (98.7%) were successfully coded. Sixty-seven percent of incidents were encompassed by the following 4 of 21 categories: prescriber ordering (22%), nursing care (15%), pharmacy (14%), and relational/communication issues (15%). Of 297 evaluable incidents, 47% did not reach the patient, 44.7% reached the patient without harm, 7.7% caused minor injury, and 0.7% caused severe injury or death. Submission rates by physicians varied between the 3 sites accounting for 1.7%, 10.7%, and 16.1% of reports. Secondary analysis identified 9 distinct scenarios that may result in major or catastrophic patient harm. CONCLUSIONS: A medical oncology-specific incident reporting taxonomy has the potential to increase our understanding of inherent risks and may lead to process improvements that improve patient safety.

Improving Home Caregiver Independence With Central Line Care for Pediatric Cancer Patients.

OBJECTIVE: Home caregivers (eg parents) of pediatric patients with cancer with external central lines (CL) must carefully maintain this device to prevent complications. No guidelines exist to support caregiver skill development, assess CL competency, follow-up after initial CL teaching, and support progress over time. We aimed to achieve >90% caregiver independence with CL care within 1 year through a family-centered quality improvement intervention. METHODS: Drivers to achieve CL care independence were identified using surveys and interviews of patient or caregivers, a multidisciplinary team with patient or family representatives, and piloting clinic return demonstrations (teach-backs). A family-centered CL care skill-learning curriculum, with a postdischarge teach-back program, was implemented using plan-do-study-act cycles. Patients or caregivers participated until independent with CL flushing. Changes included: language iterations to maximize patient or caregiver engagement, developing standardized tools for home use and for teaching and evaluating caregiver proficiency on the basis of number of nurse prompts required during the teach-back, earlier inpatient training, and clinic redesign to incorporate teach-backs into routine visits. The proportion of eligible patients whose caregiver had achieved independence in CL flushing was the outcome measure. Teach-back program participation was a process measure. Statistical process control charts tracked change over time. RESULTS: After 6 months of quality improvement intervention, >90% of eligible patients had a caregiver achieve independence with CL care. This was sustained for 30 months postintervention. Eighty-eight percent of patients (n = 181) had a caregiver participate in the teach-back program. CONCLUSION: A family-centered hands-on teach-back program can lead to caregiver independence in CL care.

The utility of procalcitonin for diagnosing bacteremia and bacterial pneumonia in hospitalized oncology patients.

PURPOSE: Procalcitonin (PCT) is an inflammatory marker elevated in bacteremia and bacterial pneumonia. We aimed to assess the real-world diagnostic accuracy of PCT in hospitalized patients with malignancy. METHODS: A retrospective cohort of 715 patients with cancer who had PCT measured during 750 admissions was analyzed. Diagnosis of bacteremia was determined using blood culture data. Diagnosis of bacterial pneumonia was based on radiographic infiltrate and/or sputum culture. PCT's performance was assessed using receiver operating characteristic (ROC) curves, sensitivity, and specificity. RESULTS: Patients had bacteremia, bacterial pneumonia, or both during 210 admissions (28%). PCT elevation above 0.5 ng/mL was significantly associated with diagnosed infection in the overall population (p < 0.0001) and in subgroups with solid tumor malignancies (p < 0.0001) and hematologic malignancies (p = 0.008). PCT was associated with infectious status in patients with any metastases, but not those with primary lung cancer, lung metastases, neuroendocrine tumors, febrile neutropenia, or history of bone marrow transplant (BMT). The area under the ROC curve for PCT in the overall population was 0.655. An ideal cutoff of 0.21 ng/mL led to a sensitivity of 60% and specificity of 59%. At cutoffs of 0.5 ng/mL and 0.05 ng/mL, PCT's sensitivity was 39% and 94%, while specificity was 79% and 17%, respectively. CONCLUSION: In this large cohort of hospitalized oncology patients, PCT elevation was associated with diagnosed bacteremia and/or bacterial pneumonia. However, specificity was limited, and PCT elevation was not associated with diagnosed infection in some subpopulations. While PCT may have some diagnostic utility for hospitalized oncology patients, values must be interpreted cautiously and considering clinical context.

Resident-Reported Impact of a Novel Oncology Curriculum for Internal Medicine Residents.

The Accreditation Council of Graduate Medical Education mandates that all internal medicine residents gain exposure to internal medicine subspecialties including hematology and oncology. While many residents meet this criterion through inpatient oncology rotations, the current structure of many inpatient oncology rotations leaves little opportunity for formal education. We therefore designed a novel oncology curriculum consisting of one-page oncology teaching sheets to increase the number, breadth, and quality of formal teaching sessions on our resident inpatient oncology services. In order to evaluate the curriculum, we conducted pre- and post-intervention surveys of residents. From these surveys, we found that 72.2% of residents used the teaching sheets on their inpatient oncology rotation and that the teaching sheets led to an increase in the number of formal oncology teaching sessions (mean 3.4 ± 2.1 post-implementation vs 2.6 ± 2.0 pre-implementation, p = 0.008), the breadth of oncology topics taught (% reporting ≥ 5 topics; 26.1% vs 16.3%, p = 0.035), the proportion of residents reporting improvement in overall oncology knowledge (80.2% vs 62.4%, p = 0.012), and the proportion of residents reporting improvement in their ability to care for patients (70.8% vs 48.9%, p = 0.013). These results demonstrate that formal oncology teaching can be improved on inpatient oncology rotations through a simple and easily replicable oncology curriculum.

Early Findings on the Use of Clinical Pathways for Management of Unwarranted Variation in Cancer Care.

Clinical pathways have the potential to improve complex clinical decision-making in cancer care. The authors implemented pathways with customized content to assist oncologists to select treatments, aiming for an on-pathway rate of 70%-85%. Treatment decisions were captured as on or off pathway, and metrics were shared monthly with users. Oncologists were categorized into quintiles based on on-pathway performance during the first 90 days of use. On-pathway rates were then calculated for days 91-360 (N = 121). Median on-pathway quintile rates varied from 50% to 100% in the initial 90-day period. During follow-up, median on-pathway rates shifted into the prespecified goal range for all groups. Clinical pathways resulted in greater uniformity in medical oncology practice. Monthly feedback about usage, familiarity with the electronic platform, and regular content updates are some factors that may influence on-pathway rates. Clinical pathways hold promise to manage unwarranted variation in cancer care.

The Future of Teaching Epidemiology.

The rapid pace of technological advancements and the corresponding societal innovations and adaptations make it difficult to predict how teaching epidemiology will look in the coming decades. We discuss changes in the teaching of epidemiology that are currently unfolding. First, typical epidemiology curricula often lack formal instruction in important components of causal thinking, such as the formulation of well-defined research questions. We address gaps related to causal thinking, communication about our science, and interpretation of study results, and we make suggestions of specific content to close such gaps. Second, digital technology increasingly influences epidemiology instruction. We discuss classroom and online teaching modalities in terms of challenges and advantages.

Role of Adverse Events in Unscheduled Hospitalizations Among Patients With Solid Tumors Who Receive Medical Oncology Treatment.

PURPOSE: The development of strategies to prevent or mitigate cancer treatment-related adverse events (AEs) is necessary to improve patient experience, safety, and cost containment. To develop a strategy to easily identify and mitigate AEs, we sought to understand the frequency and severity of those that resulted in hospitalizations. METHODS: We retrospectively characterized hospitalizations of ambulatory adult patients with solid tumor cancers within 30 days of chemotherapy administration using medical record data abstraction. Hospitalizations were categorized as caused by cancer symptoms, a noncancer medical condition, or a medical oncology treatment-related AE. Severity of the treatment-related AE hospitalization was rated using the National Patient Safety Agency risk assessment matrix scale. RESULTS: Between May and October 2016, 116 patients experienced 197 hospitalizations (per-patient mean, 1.7 AEs; range, 1 to 7 AEs). Sixty-six percent (n = 130) of hospitalizations were related to cancer symptoms, whereas 19.3% (n = 38) were treatment-related AE hospitalizations. The median length of stay of hospitalizations that resulted from an AE was 6 days (interquartile range, 3 to 9 days), and 36.8% had more than 1 AE. GI symptoms accounted for 48.1% of AEs, and neutropenic fever accounted for 11.1%. Sixty-one percent of treatment-related AE hospitalizations were characterized as moderate severity. CONCLUSION: Hospitalizations in patients with solid tumors as a direct result of their medical oncology care treatment are not uncommon. These findings argue for novel approaches, such as automated trigger tools, to identify and manage complications of medical oncology treatment before hospitalization is needed. Improved outpatient management of cancer symptoms may have a dramatic impact on hospitalizations for patients with cancer.

Perceived racism and incident diabetes in the Black Women's Health Study.

AIMS/HYPOTHESIS: Our aim was to assess the association of perceived racism with type 2 diabetes, and the possible mediating influence of diet and BMI. METHODS: The Black Women's Health Study, a follow-up of 59,000 African-American women, began in 1995. Over 16 years 5344 incident cases of diabetes occurred during 576,577 person-years. Cox proportional hazards models were used to estimated HRs and 95% CIs for categories of 'everyday racism' (interpersonal racism in daily life) and 'lifetime racism' (reporting ever treated unfairly due to race with respect to police, housing or work) and incident type 2 diabetes. Models were adjusted for age, questionnaire cycle, marital status, socioeconomic status, education, family history of diabetes, physical activity, alcohol use and smoking status, with and without inclusion of terms for dietary patterns and adult BMI. RESULTS: Compared with women in the lowest quartile of exposure, women in the highest quartile of exposure to everyday racism had a 31% increased risk of diabetes (HR 1.31; 95% CI 1.20, 1.42) and women with the highest exposure to lifetime racism had a 16% increased risk (HR 1.16; 95% CI 1.05, 1.27). Mediation analysis estimated that BMI accounted for half of the association between either the everyday or lifetime racism measure and incident diabetes. CONCLUSIONS/INTERPRETATION: Perceived everyday and lifetime racism were associated with increased risk of type 2 diabetes in this cohort of African-American women and appear to be at least partly mediated by BMI.

Characteristics Associated with In-Hospital Death among Commercially Insured Decedents with Cancer.

BACKGROUND: A majority of patients with poor-prognosis cancer express a preference for in-home death; however, in-hospital deaths are common. OBJECTIVE: We sought to identify characteristics associated with in-hospital death. DESIGN: Case series. SETTING/SUBJECTS: Commercially insured patients with cancer who died between July 2010 and December 2013 and who had at least two outpatient visits at a tertiary cancer center during the last six months of life. MEASUREMENTS: Patient characteristics, healthcare utilization, and in-hospital death (primary outcome) were ascertained from institutional records and healthcare claims. Bivariate and multivariable analyses were used to evaluate the association of in-hospital death with patient characteristics and end-of-life outcome measures. RESULTS: We identified 904 decedents, with a median age of 59 years at death. In-hospital death was observed in 254 patients (28%), including 110 (12%) who died in an intensive care unit. Hematologic malignancy was associated with a 2.57 times increased risk of in-hospital death (95% confidence interval [CI] 1.91-3.45, p < 0.001), and nonenrollment in hospice was associated with a 14.5 times increased risk of in-hospital death (95% CI 9.81-21.4, p < 0.001). Time from cancer diagnosis to death was also associated with in-hospital death (p = 0.003), with the greatest risk among patients dying within six months of cancer diagnosis. All significant associations persisted in multivariable analyses that were adjusted for baseline characteristics. CONCLUSIONS: In-hospital deaths are common among commercially insured cancer patients. Patients with hematologic malignancy and patients who die without receiving hospice services have a substantially higher incidence of in-hospital death.

Novel Data Sharing Between a Comprehensive Cancer Center and a Private Payer to Better Understand Care at the End of Life.

CONTEXT: Understanding end-of-life (EOL) care patterns is a prerequisite to improving the experience for cancer patients. EOL measures endorsed by the National Quality Forum (NQF) have been examined in older patients using Medicare claims. OBJECTIVES: To evaluate EOL care for patients treated at a comprehensive cancer center, using private payer claims data. METHODS: A retrospective cohort study was conducted of Dana-Farber Cancer Institute (DFCI) patients who died between July 2010 and December 2012, and were insured by Blue Cross Blue Shield of Massachusetts. Primary data sources included Blue Cross Blue Shield of Massachusetts claims information and DFCI administrative data. We assessed NQF-endorsed measures of EOL care related to emergency department visits, hospitalizations, and intensive care unit admissions in the last 30 days, chemotherapy in the last 14 days, hospice stay, and death in an acute care setting. Patterns of care by cancer type and service location were determined. RESULTS: Among 674 patients (mean age 58 years), event rates for NQF-endorsed EOL measures were similar to those reported using Medicare claims. Decedents with hematologic malignancies received significantly more intensive care and were less likely to have enrolled in hospice, compared to decedents with solid tumors. Thirty to 45% of EOL events occurred outside of DFCI and its affiliated hospitals. CONCLUSION: Data sharing between a private payer and a large cancer center proved feasible and informative. High rates of hospital service use outside of our sites of care were unexpected. The findings suggest opportunities to better manage care at the end of life.

Understanding Oral Chemotherapy Prescribing Patterns at the End of Life at a Comprehensive Cancer Center: Analysis of a Massachusetts Payer Claims Database.

PURPOSE: Receipt of chemotherapy in the last 14 days of life is a measure of potential overuse of care. Specific measures defining appropriate end-of-life use of oral agents have not yet been described, and little is known about prescribing patterns. METHODS: We conducted an exploratory analysis of 371 patients at Dana-Farber Cancer Institute who were covered by the Blue Cross Blue Shield of Massachusetts pharmacy benefit and died during 2012 to 2013. We analyzed processed claims as a surrogate for chemotherapy administration. We compared oral with parenteral chemotherapy claims in the last 6 months of life. RESULTS: In the last 6 months of life, 294 patients (79%) had chemotherapy claims, including 81 (22%) prescribed an oral agent; 20 patients had claims for oral chemotherapy in the last 30 days of life. For eight patients (40%), this was the initial start of that oral agent. In the last 14 days of life, only 23 patients had chemotherapy claims, including six patients prescribed an oral agent. CONCLUSION: The collection of oral chemotherapy use data through insurance claims was feasible. Processed claims for chemotherapy, including oral, sharply declined during the last 30 days of life, consistent with a shift to palliative management. These results highlight the need for a more comprehensive analysis of oral chemotherapy prescribing patterns and development of specific measures to define the appropriate use of oral chemotherapy at the end of life.

Perceptions of medical errors in cancer care: an analysis of how the news media describe sentinel events.

OBJECTIVE: To analyze the print news media's coverage of sentinel events involving cancer patients. METHODS: Using LexisNexis, we identified English-language newspaper articles covering medical errors in cancer care between January 1, 2000, and December 31, 2010. Articles were coded for 3 major themes using a standardized abstraction instrument: narrative statements and point of view most prominently represented, attribution of blame, and orientation toward patient safety. We also abstracted country where the newspaper was published, type of error event, and extent of patient harm. RESULTS: We analyzed 64 articles from 37 print newspaper syndications that circulated in 6 countries/regions. Reports of medical errors rarely were framed from the point of view of a safety expert or the responsible clinician (13% and 3%, respectively) compared with the patient and legal points of view (both 30%). Articles held individual clinicians (41%) and hospital systems (28%) responsible for most errors. Four in 10 articles failed to present medical errors as "systems" problems. Article perspective varied considerably by country, with 53% of articles from the UK and 63% from Australia and New Zealand judged as negatively slanted compared with 14% in the United States and Canada. CONCLUSIONS: In reports of medical errors involving cancer patients, the news media regularly blame individual clinicians for mistakes and fail to present a systems-based understanding of these events.

Impact of a single nucleotide polymorphism upstream of the IL28B gene in patients positive for anti-HCV antibody in an HCV hyperendemic area in Japan.

The influence of genetic variation at the interleukin-28B (IL28B) locus on the natural course of hepatitis C virus (HCV) infection has not been fully investigated. The goal of this study was to examine whether an IL28B polymorphism (rs8099917) is associated with natural clearance of HCV and with disease parameters of HCV infection in an HCV hyperendemic area of Japan. The patients were 502 anti-HCV antibody-positive residents who participated in liver disease screening program from 2002 to 2004. Patients who underwent interferon-based therapy or had hepatocellular carcinoma were excluded. Of these patients, 149 were negative for HCV RNA (prior infection) and 353 were positive for HCV RNA or HCV core antigen (HCV carriers). In multivariate analysis, the IL28B TT genotype was a predictor for prior HCV infection. In addition, nine of the patients with prior HCV infection were positive for anti-HCV antibody with positive for HCV core antigen or HCV RNA before 2001, and these nine patients all had the IL28B TT genotype. Furthermore, the IL28B TT genotype was associated independently with higher HCV core antigen levels in HCV carriers. In contrast, the IL28B genotype did not affect the biochemical markers, such as alanine aminotransferase, hepatic fibrosis markers, and α-fetoprotein, and the degree of hepatic fibrosis assessed by transient elastography in HCV carriers. We concluded that IL28B polymorphism (TT genotype) is associated with spontaneous clearance of HCV and conversely with high viral loads in HCV carriers. In contrast, the IL28B genotype does not affect disease progression such as hepatic fibrosis.

Impact of antibody to hepatitis B core antigen on the clinical course of hepatitis C virus carriers in a hyperendemic area in Japan: A community-based cohort study.

AIM: Subjects positive for antibody to hepatitis B core antigen (HBcAb) and negative for hepatitis B surface antigen (HBsAg) are considered to have occult hepatitis B virus (HBV) infection. The aim of this study was to determine the impact of occult HBV infection on aggravation of the clinical course in hepatitis C virus (HCV) carriers. METHODS: A prospective cohort study was performed in 400 subjects who were positive for anti-HCV antibody and negative for HBsAg. Among these subjects, 263 were HCV core antigen positive or HCV RNA positive (HCV carriers). We examined whether the presence of HBcAb affected the clinical course in these HCV carriers from 1996-2005. RESULTS: The HBcAb positive rates were 53.6% and 52.6% in HCV carriers and HCV RNA negative subjects, respectively. There were no differences in the incidence of hepatocellular carcinoma (HCC) and cumulative mortality associated with liver-related death between HCV carriers who were positive and negative for HBcAb. In multivariate analysis, age (≥65 years) and alanine aminotransferase level (≥31 IU/L) emerged as independent risk factors for HCC development and liver-related death, but the HBcAb status was not a risk factor. In addition, increased serum hepatic fibrosis markers (measured from 2001-2004) were not associated with HBcAb status. CONCLUSION: In our cohort study, the presence of HBcAb had no impact on HCC development, liver-related death and hepatic fibrosis markers in HCV carriers. Thus, our results indicate that occult HBV infection has no impact on the clinical course in HCV carriers.

Predictors of active injection drug use in a cohort of patients infected with hepatitis C virus.

OBJECTIVES: We investigated potential risk factors for active injection drug use (IDU) in an inner-city cohort of patients infected with hepatitis C virus (HCV). METHODS: We used log-binomial regression to identify factors independently associated with active IDU during the first 3 years of follow-up for the 289 participants who reported ever having injected drugs at baseline. RESULTS: Overall, 142 (49.1%) of the 289 participants reported active IDU at some point during the follow-up period. In a multivariate model, being unemployed (prevalence ratio [PR] = 1.93; 95% confidence interval [CI] = 1.24, 3.03) and hazardous alcohol drinking (PR = 1.67; 95% CI = 1.34, 2.08) were associated with active IDU. Smoking was associated with IDU but this association was not statistically significant. Patients with all 3 of those factors were 3 times as likely to report IDU during follow-up as those with 0 or 1 factor (PR = 3.3; 95% CI = 2.2, 4.9). Neither HIV coinfection nor history of psychiatric disease was independently associated with active IDU. CONCLUSIONS: Optimal treatment of persons with HCV infection will require attention to unemployment, alcohol use, and smoking in conjunction with IDU treatment and prevention.

Comparing clinicians' use of an anticoagulation management service and usual care in ambulatory oncology.

PURPOSE: There is no consensus in the oncology community about the optimal model for anticoagulation management of ambulatory cancer patients. To understand oncologists' preferences regarding anticoagulation management, we compared the characteristics of patients referred to an oncology-oriented anticoagulation management service with "usual care" patients managed by the patient's primary oncologist. METHODS: We performed a retrospective medical record review of ambulatory oncology patients' anticoagulation care at a comprehensive cancer center. We examined the characteristics of 33 patients anticoagulated before implementation of a dedicated oncology anticoagulation management service. We compared this group with 33 patients managed by the anticoagulation management service and with 39 usual care patients managed by the primary oncologist after the anticoagulation management service was created. We also examined differences in laboratory test utilization, time in the therapeutic range (for patients anticoagulated with warfarin), and anticoagulation-related adverse events during a 3-month assessment period. RESULTS: Anticoagulation management service patients were more likely to be treated for hematologic malignancies, use erythropoietin stimulating agents, and require warfarin management for previous venous thromboembolic disease compared to usual care patients. In contrast, oncologists were more likely to manage anticoagulation care of patients with advanced solid tumors undergoing active chemotherapy. Anticoagulation management service and usual care patients on warfarin therapy had comparable time in the therapeutic range and complication rates. CONCLUSION: Oncologists selectively referred patients to the anticoagulation management service. Anticoagulation management service patients' warfarin control and complication rates were comparable to care provided by the primary oncologist, suggesting that an oncology-specific anticoagulation management service may be a feasible and effective option for anticoagulation management of ambulatory oncology patients.

Factors associated with pain among ambulatory patients with cancer with advanced disease at a comprehensive cancer center.

PURPOSE: The prevalence and severity of pain have not been well described among oncology patients in ambulatory care. To better understand the burden of pain among patients with advanced cancer, we examined the prevalence of pain reported during office and treatment visits. METHODS: A retrospective study of 4,014 patients with advanced disease (stage 4 at diagnosis or metastatic progression) who completed an ambulatory visit between 2004 and 2006 was conducted at a comprehensive cancer center in Boston, Massachusetts. RESULTS: At their first visit during the study period, 74% of patients reported no pain (0 score); 12%, low pain (1 to 3 score); 9%, moderate pain (4 to 6 score); and 5%, severe pain (7 to 10 score). The prevalence of pain was highest among patients who were younger than 60 years of age, were nonwhite, did not speak English as their primary language, or were covered by Medicaid, received free care, or paid their own health care costs. Patients with thoracic, breast, and head and neck cancers had higher pain scores than those with other diseases. Pain was reported more frequently among patients whose diagnosis or metastatic progression occurred less than 3 months before the reported pain score. In multivariable regression analysis, age, race, cancer type, and time since diagnosis/progression were identified as important factors associated with severe pain. CONCLUSION: Younger age, minority race, and recent onset of advanced disease are associated with severe pain among patients with cancer. Recognizing these high-risk groups could inform targeted interventions to address pain care in ambulatory patients with advanced cancer.

Incidence of severe pain in newly diagnosed ambulatory patients with stage IV cancer.

BACKGROUND: Pain is common among cancer patients. OBJECTIVE: To characterize the incidence of severe pain among newly diagnosed patients with stage IV cancer in ambulatory care. METHODS: A retrospective cohort of 505 ambulatory oncology patients with newly diagnosed stage IV solid tumours at a comprehensive cancer centre (Dana-Farber Cancer Institute, Boston, Massachusetts, USA) was followed from January 1, 2004, to December 31, 2006. Pain intensity scores were extracted from electronic medical records. The incidence of severe pain was calculated using the maximum monthly pain scores reported at outpatient visits. RESULTS: Of the 505 patients included in the present study, 340 (67.3%) were pain-free at the initial visit, 90 (17.8%) experienced mild pain, 48 (9.5%) experienced moderate pain and 27 (5.4%) experienced severe pain. At least one episode of severe pain within one year of diagnosis was reported by 29.1% of patients. Patients with head and neck, gastrointestinal and thoracic malignancies were more likely to experience severe pain compared with patients with other types of cancer (52.6%, 33.9% and 30.5%, respectively). In the multivariable model, patients whose primary language was not English (OR 2.90 [95% CI 1.08 to 7.80]), patients who reported severe pain at the initial visit (OR 9.30 [95% CI 3.72 to 23.23]) and patients with head and neck (OR 10.17 [95% CI 2.87 to 36.00]) or gastrointestinal (OR 4.05 [95% CI 1.23 to 13.35]) cancers were more likely to report severe pain in the following year. CONCLUSIONS: The incidence of severe pain was high in ambulatory patients with newly diagnosed stage IV cancer.

Assessing the quality of pain care in ambulatory patients with advanced stage cancer.

CONTEXT: Pain is common among patients with advanced cancer despite the dissemination of clinical pain care guidelines. OBJECTIVES: We sought to assess the quality of pain care among patients with advanced disease. METHODS: We reviewed the records of 85 adult ambulatory patients with advanced breast, lung, and gastrointestinal cancer treated in 2004-2006. Patients' screening pain intensity scores were at least 7 of 10. Nurse reviewers completed medical record reviews of care rendered at the index visit and over the subsequent 30 days based on the 2004 National Comprehensive Cancer Network pain guideline. An expert panel then rated the quality of the evaluation, treatment, and overall pain care. We used a multivariable model to analyze guideline compliance and resolution of severe pain. RESULTS: Among advanced cancer patients with severe pain, clinicians adjusted pain medications only half the time and made few timely referrals for pain-related consultations. By 30 days after the index visit, 34% of patients continued to report severe pain. The expert panel judged the overall quality of pain care as "fair" or "poor" in about two-thirds of cases because more timely and effective intervention could have reduced the severity and duration of pain. Resolution of severe pain was associated with adjustment of pain medications at the index visit (adjusted odds ratio 3.8, 95% CI 1.3-10.6). CONCLUSION: There is room for improvement in the pain care of patients with advanced cancer. Additional research is needed to understand the reasons for poor performance.