Exploring Medicaid claims data to understand predictors of healthcare utilization and mortality for Medicaid individuals with or without a diagnosis of lung cancer: a feasibility study

Exploring Medicaid claims data to understand predictors of healthcare utilization and mortality... Abstract Health disparities in low-income populations complicate care for at-risk individuals or those diagnosed with lung cancer and may influence their patterns of healthcare utilization. The purpose of this study is to examine whether age, sex, provider’s affiliation, Medicare dual eligibility, and number of comorbidities can predict healthcare utilization, as well as to examine factors influencing mortality in lung biopsy patients. A retrospective review of de-identified Medicaid claims of adults having a lung biopsy in 2013 resulted in classification into lung cancer and non–lung cancer cases based on a lung cancer diagnostic code within 30 days after biopsy. Biopsy cases were further divided by whether or not the provider’s institution was accredited by the Commission on Cancer (CoC). Inpatient (IP), outpatient (OP), and emergency department (ED) utilization was followed from initial date of biopsy through 2015, or to the earliest date of death, disenrollment, or study end for both groups. The result of Cox proportional hazards regression model indicated that age and the number of comorbidities significantly predicted OP use and the number of comorbidities significantly predicted ED use in patients with lung cancer. However, for non–lung cancer patients, only the number of comorbidities significantly predicted IP and ED uses. Furthermore, for patients with lung cancer, the significant factors of mortality included IP use per month and the number of comorbidities. Patients with lung cancer who received a lung biopsy by a CoC-accredited organization had a longer time of survival from the biopsy event. Our findings suggest that understanding predictors of healthcare utilization and mortality may create opportunities to improve health and quality of life through better healthcare coordination. Implications Research: Future research is needed to examine healthcare utilization patterns with recommended evidence-based guidelines for patients with lung cancer on Medicaid. Practice: Understanding healthcare utilization patterns for low-income patients with lung cancer can provide opportunities to identify healthcare disparities in access, cost, and quality and can further support the development of appropriate and effective interventions to improve overall quality of care. Policy: Health organizations and policy makers may need to take new approaches to promote data sharing to create robust databases to facilitate lung cancer research that improves quality of care. INTRODUCTION Limited access to health resources and poor coordination of healthcare in low-income populations is associated with underuse of cancer screening, higher risk for a late diagnosis of cancer, increased symptoms related to a cancer diagnosis and comorbid disease, decreased quality of life (QOL), and overall survival [1, 2]. Care coordination is the organization of patient care activities including access to resources and exchange of critical information to facilitate the appropriate delivery of healthcare services [3]. Several factors influence health outcomes including lack of information at point of care, as well as poor communication between primary and specialty care relating to inpatient (IP) stays, outpatient (OP) visits, and emergency department (ED) visits. Integrating coordination programs continues to be challenging at the provider level as care delivery processes and plans can be both time-consuming and financially burdensome [4, 5]. Although there have been initiatives over the past decade to develop cancer survivorship care plans (SCPs) [6], there is little known regarding healthcare utilization and care coordination for lung cancer. Lung cancer is the most common cause of cancer death among the four major cancers [7]. Cancer care disparities are complex and not well understood; some of the reasons for this are patient related including their health behavior, healthcare provider, as well as their varying capacity to interact with the healthcare system [5, 8]. Attention to lung cancer has increased with new advances and collaborative work among clinicians and scientists; however, health disparities still exist in managing health-directed treatment, comorbidities, QOL, and symptom management [5, 8–11]. Health-related QOL for lung cancer survivors are not equitable to age-matched survivors of other cancers [11]. Utilization of appropriate healthcare services may help to address such issues as compromised lung capacity, overwhelming symptom burden, and treatment-related morbidity affecting QOL [11, 12]. In addition, the burden of disease varies among populations based on insurance status [5]. Differences in groups are evident in mortality, access to care, and delivery of care. Patients with lung cancer with Medicaid insurance have higher incidence rates, poorer outcomes with more advanced disease at diagnosis, and poorer survival. Patients with Medicaid are more likely than others to die during the month they were diagnosed and are less likely to receive surgery, radiation, or treatment at centers managing a high volume of patients [5]. The changing landscape of the Affordable Care Act has created angst among many health service providers regarding access to specialized cancer care especially for Medicaid beneficiaries [13–15]. Gaps and restrictions in Medicaid coverage can create barriers facilitating continued disparities in access to care, narrowed provider networks to provide specialized cancer services, and meeting evidence-based guidelines for quality care [15]. Quality measures and national benchmarks are essential tools to ensure evidence-based guideline care for patients with cancer [15–17]. The Commission on Cancer (CoC) was established by the American College of Surgeons to provide an accredited process to evaluate oncologic outcomes and the QOL for patients with cancer and to ensure continuity of care and access to cancer resources and services [18–20]. CoC accreditation is based on compliance and adherence to recommended guidelines for cancer, establishing the benchmark for delivery of specialized care, and addressing health disparities and barriers to care which is important when evaluating health service utilization and quality monitoring of individuals enrolled in Medicaid. Low-income and younger lung cancer survivors may not utilize appropriate healthcare services and therefore may not be meeting national evidence-based recommendations for disease management of prevention, screening, treatment, supportive, and survivorship care [1, 10]. Reducing health disparities underscores the critical need for understanding health utilization and lung cancer care in a younger, diverse, low-income, understudied, and higher risk population such as those participating in a Medicaid insurance program. Purpose The purpose of this study is to examine whether age, sex, provider’s affiliation, Medicare dual eligibility, and the number of comorbidities can predict healthcare utilization, as well as to examine factors influencing mortality in lung biopsy patients. METHODS Study design The population, variables, and analysis selected address the gaps in healthcare utilization in lung biopsy cases. The analysis compares individuals with a biopsy followed by a lung cancer diagnosis to individuals with biopsy without a cancer diagnosis. We conducted a retrospective review of de-identified Medicaid claims of adults in eight counties in Western New York (Erie, Niagara, Orleans, Genesee, Wyoming, Allegany, Cattaraugus, and Chautauqua) who had a lung biopsy between January 1 and December 31, 2013. Cases were followed from initial date of biopsy through December 31, 2015 or to the earliest date of death, disenrollment, or study end for both groups. We stratified both lung cancer and noncancer cases by initial biopsy in CoC (CoC accreditation status) and non-CoC accreditation. We captured data on IP, OP, and ED visits and mortality. The University at Buffalo Institutional Review Board reviewed the protocol for de-identification and approved it as not human subject research. Data source The secondary data analysis utilized existing de-identified demographics and claims data extracted from the Medicaid Data Warehouse (MDW). The claims data include the same research ID, a flag for type of claim (e.g., IP, ED, OP), dates-of-service converted to interval from a random date, a flag for the provider of a specific service, and the first five diagnosis codes. As the MDW contains protected health information, analysts extracted Health Insurance Portability and Accountability Act of 1996 (HIPAA) compliant demographic and claims data and assigned each case a unique research ID. The demographic database consists of research ID, age, sex, county of residence, Medicare eligibility, and a flag for death within a year. Measures We used the International Classification of Disease based on the Healthcare Cost and Utilization Project (H-CUP) definition for lung cancer and major chronic conditions. H-CUP’s Clinical Classification Software (CCS) aggregates diagnosis codes into conditions [21]. Chronic conditions were divided into two groups based on the number of comorbidities (one to two comorbidities, and three to eight comorbidities). See Table 1 for details of codes included in cancer and chronic disease groups. Table 1 International Classification of Disease (ICD)-9 and Current Procedural Terminology (CPT) codes used to identify patients with cancer of bronchus and lung Domain  CCS codes for ICD-9  Descriptions  Chronic disease base  49, 50, 186 (DM); 66, 67 (SA); 65, 68, 69, 70, 71, 72, 73, 74, 75 (MH); 98, 183 (HTN); 99a, 158 (CKD); 103, 108 (HF); 100, 101 (CAD); 127 (COPD); 128 (ASTH)  Diabetes (DM); substance abuse (SA); mental health (MH); hypertension (HTN); chronic kidney disease (CKD); heart failure (HF); coronary artery disease (CAD); chronic obstructive pulmonary disease (COPD); asthma (ASTH)  Cancer  13; 14; 15; 17; 19,20; 22; 24; 25; 27; 29; 30; 33; 36  Stomach; colon; rectal; pancreatic; lung; melanoma; breast; uterine; ovarian; prostate; bladder; kidney; thyroid  Bronchus and lung cancer  1622  MALIG NEO MAIN BRONCHUS    1623  MAL NEO UPPER LOBE LUNG    1624  MAL NEO MIDDLE LOBE LUNG    1625  MAL NEO LOWER LOBE LUNG    1628  MAL NEO BRONCH/LUNG NEC    1629  MAL NEO BRONCH/LUNG NOS    20921  MAL CARCINOID BRONC/LUNG    2312  CA IN SITU BRONCHUS/LUNG    CPT codes for lung biopsy  Description  Codes for lung biopsy  31628  Bronchoscopy with transbronchial lung biopsy(ies) single lobe    31629  Bronchoscopy with transbronchial needle aspiration biopsy(ies) trachea, main stem and/or lobar bronchus(i)    31632  Bronchoscopy with transbronchial lung biopsy(ies), each additional lobe    31633  Bronchoscopy with transbronchial needle aspiration biopsy(ies), each additional lobe    32096  Thoracotomy with diagnostic biopsy(ies) of lung infiltrates (e.g., wedge, incisional), unilateral    32097  Thoracotomy with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32098  Thoracotomy with biopsy(ies) of pleura    32400  Biopsy pleura; percutaneous needle    32405  Biopsy lung or mediastinum, percutaneous needle    32606  Thoracoscopy, mediastinal space, with biopsy    32607  Thoracoscopy; with diagnostic biopsy(ies) of lung infiltrate(s) (e.g., wedge, incisional), unilateral    32608  Thoracoscopy; with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32609  Thoracoscopy; with biopsy(ies) of pleura    39000  Open biopsy of mediastinum    39010  Open biopsy of mediastinum  Domain  CCS codes for ICD-9  Descriptions  Chronic disease base  49, 50, 186 (DM); 66, 67 (SA); 65, 68, 69, 70, 71, 72, 73, 74, 75 (MH); 98, 183 (HTN); 99a, 158 (CKD); 103, 108 (HF); 100, 101 (CAD); 127 (COPD); 128 (ASTH)  Diabetes (DM); substance abuse (SA); mental health (MH); hypertension (HTN); chronic kidney disease (CKD); heart failure (HF); coronary artery disease (CAD); chronic obstructive pulmonary disease (COPD); asthma (ASTH)  Cancer  13; 14; 15; 17; 19,20; 22; 24; 25; 27; 29; 30; 33; 36  Stomach; colon; rectal; pancreatic; lung; melanoma; breast; uterine; ovarian; prostate; bladder; kidney; thyroid  Bronchus and lung cancer  1622  MALIG NEO MAIN BRONCHUS    1623  MAL NEO UPPER LOBE LUNG    1624  MAL NEO MIDDLE LOBE LUNG    1625  MAL NEO LOWER LOBE LUNG    1628  MAL NEO BRONCH/LUNG NEC    1629  MAL NEO BRONCH/LUNG NOS    20921  MAL CARCINOID BRONC/LUNG    2312  CA IN SITU BRONCHUS/LUNG    CPT codes for lung biopsy  Description  Codes for lung biopsy  31628  Bronchoscopy with transbronchial lung biopsy(ies) single lobe    31629  Bronchoscopy with transbronchial needle aspiration biopsy(ies) trachea, main stem and/or lobar bronchus(i)    31632  Bronchoscopy with transbronchial lung biopsy(ies), each additional lobe    31633  Bronchoscopy with transbronchial needle aspiration biopsy(ies), each additional lobe    32096  Thoracotomy with diagnostic biopsy(ies) of lung infiltrates (e.g., wedge, incisional), unilateral    32097  Thoracotomy with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32098  Thoracotomy with biopsy(ies) of pleura    32400  Biopsy pleura; percutaneous needle    32405  Biopsy lung or mediastinum, percutaneous needle    32606  Thoracoscopy, mediastinal space, with biopsy    32607  Thoracoscopy; with diagnostic biopsy(ies) of lung infiltrate(s) (e.g., wedge, incisional), unilateral    32608  Thoracoscopy; with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32609  Thoracoscopy; with biopsy(ies) of pleura    39000  Open biopsy of mediastinum    39010  Open biopsy of mediastinum  CCS Clinical Classification Software. aCodes for HTN, CKD, and HF. View Large Table 1 International Classification of Disease (ICD)-9 and Current Procedural Terminology (CPT) codes used to identify patients with cancer of bronchus and lung Domain  CCS codes for ICD-9  Descriptions  Chronic disease base  49, 50, 186 (DM); 66, 67 (SA); 65, 68, 69, 70, 71, 72, 73, 74, 75 (MH); 98, 183 (HTN); 99a, 158 (CKD); 103, 108 (HF); 100, 101 (CAD); 127 (COPD); 128 (ASTH)  Diabetes (DM); substance abuse (SA); mental health (MH); hypertension (HTN); chronic kidney disease (CKD); heart failure (HF); coronary artery disease (CAD); chronic obstructive pulmonary disease (COPD); asthma (ASTH)  Cancer  13; 14; 15; 17; 19,20; 22; 24; 25; 27; 29; 30; 33; 36  Stomach; colon; rectal; pancreatic; lung; melanoma; breast; uterine; ovarian; prostate; bladder; kidney; thyroid  Bronchus and lung cancer  1622  MALIG NEO MAIN BRONCHUS    1623  MAL NEO UPPER LOBE LUNG    1624  MAL NEO MIDDLE LOBE LUNG    1625  MAL NEO LOWER LOBE LUNG    1628  MAL NEO BRONCH/LUNG NEC    1629  MAL NEO BRONCH/LUNG NOS    20921  MAL CARCINOID BRONC/LUNG    2312  CA IN SITU BRONCHUS/LUNG    CPT codes for lung biopsy  Description  Codes for lung biopsy  31628  Bronchoscopy with transbronchial lung biopsy(ies) single lobe    31629  Bronchoscopy with transbronchial needle aspiration biopsy(ies) trachea, main stem and/or lobar bronchus(i)    31632  Bronchoscopy with transbronchial lung biopsy(ies), each additional lobe    31633  Bronchoscopy with transbronchial needle aspiration biopsy(ies), each additional lobe    32096  Thoracotomy with diagnostic biopsy(ies) of lung infiltrates (e.g., wedge, incisional), unilateral    32097  Thoracotomy with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32098  Thoracotomy with biopsy(ies) of pleura    32400  Biopsy pleura; percutaneous needle    32405  Biopsy lung or mediastinum, percutaneous needle    32606  Thoracoscopy, mediastinal space, with biopsy    32607  Thoracoscopy; with diagnostic biopsy(ies) of lung infiltrate(s) (e.g., wedge, incisional), unilateral    32608  Thoracoscopy; with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32609  Thoracoscopy; with biopsy(ies) of pleura    39000  Open biopsy of mediastinum    39010  Open biopsy of mediastinum  Domain  CCS codes for ICD-9  Descriptions  Chronic disease base  49, 50, 186 (DM); 66, 67 (SA); 65, 68, 69, 70, 71, 72, 73, 74, 75 (MH); 98, 183 (HTN); 99a, 158 (CKD); 103, 108 (HF); 100, 101 (CAD); 127 (COPD); 128 (ASTH)  Diabetes (DM); substance abuse (SA); mental health (MH); hypertension (HTN); chronic kidney disease (CKD); heart failure (HF); coronary artery disease (CAD); chronic obstructive pulmonary disease (COPD); asthma (ASTH)  Cancer  13; 14; 15; 17; 19,20; 22; 24; 25; 27; 29; 30; 33; 36  Stomach; colon; rectal; pancreatic; lung; melanoma; breast; uterine; ovarian; prostate; bladder; kidney; thyroid  Bronchus and lung cancer  1622  MALIG NEO MAIN BRONCHUS    1623  MAL NEO UPPER LOBE LUNG    1624  MAL NEO MIDDLE LOBE LUNG    1625  MAL NEO LOWER LOBE LUNG    1628  MAL NEO BRONCH/LUNG NEC    1629  MAL NEO BRONCH/LUNG NOS    20921  MAL CARCINOID BRONC/LUNG    2312  CA IN SITU BRONCHUS/LUNG    CPT codes for lung biopsy  Description  Codes for lung biopsy  31628  Bronchoscopy with transbronchial lung biopsy(ies) single lobe    31629  Bronchoscopy with transbronchial needle aspiration biopsy(ies) trachea, main stem and/or lobar bronchus(i)    31632  Bronchoscopy with transbronchial lung biopsy(ies), each additional lobe    31633  Bronchoscopy with transbronchial needle aspiration biopsy(ies), each additional lobe    32096  Thoracotomy with diagnostic biopsy(ies) of lung infiltrates (e.g., wedge, incisional), unilateral    32097  Thoracotomy with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32098  Thoracotomy with biopsy(ies) of pleura    32400  Biopsy pleura; percutaneous needle    32405  Biopsy lung or mediastinum, percutaneous needle    32606  Thoracoscopy, mediastinal space, with biopsy    32607  Thoracoscopy; with diagnostic biopsy(ies) of lung infiltrate(s) (e.g., wedge, incisional), unilateral    32608  Thoracoscopy; with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32609  Thoracoscopy; with biopsy(ies) of pleura    39000  Open biopsy of mediastinum    39010  Open biopsy of mediastinum  CCS Clinical Classification Software. aCodes for HTN, CKD, and HF. View Large We identified the presence or absence of lung cancer by examining de-identified Medicaid claims of adults (≥18 to ≤64 years of age) having a lung biopsy in 2013. Specific Current Procedural Terminology (CPT) codes, as shown in Table 1, were used to identify patients admitted for a lung biopsy. As cancer stage and initial date of primary diagnosis were not available, a proxy date of lung cancer diagnosis was created using a 30-day window post diagnostic lung biopsy. This procedure created two groups: lung cancer cases and non–lung cancer cases. This approach provided a practical solution for identifying individuals with a lung cancer diagnosis and for identifying lung biopsy patients who may be a population at risk for lung cancer. This approach was particularly useful for facilitating analysis when specific cancer registries or detailed patient data are not available. In addition, as Medicaid data only include date of death, we used all-cause mortality rate as a measure of the number of deaths in this study. Utilization was based on data definitions for IP, ED, and OP events. IP events were identified for all claims with a hospital listed as the place of service and eliminating all nonadmission events, such as, emergency care, laboratory and imaging studies, and other clinic visits. ED and OP events were similarly identified using evaluation and management CPT codes specific to those settings. Utilization for each person with a biopsy was tracked from the initial date of biopsy to the earliest of death, disenrollment, or end of study for both groups. We created a flag for the provider to indicate whether or not the biopsy was completed in a CoC-accredited organization. Data analysis The unit of analysis was the individual Medicaid recipient who represented a lung biopsy case. Descriptive analyses summarized demographic characteristics. The data are presented as means and standard deviations (SD) for continuous variables, as well as frequencies and percentages for categorical variables. Two group comparisons were conducted to evaluate the overall value of each variable for a significant difference in cancer and noncancer groups. For each of the two groups (i.e., cancer and noncancer), we conducted a descriptive analysis for eight variables and obtained a mean and an SD values for numeric (continuous) variables and the number of occurrence and percentage for categorical (factor) variables. For categorical or binary variables, gender, CoC, and MCARE, the cell values in the cancer and noncancer columns are the number of events and corresponding percentage, and the chi-square test was used to make the comparison of categorical variables. For numeric variables, the cell values in the cancer and noncancer columns are the mean and SD of the variables, and we used t test to test the differences between these numeric variables. Multiple linear regressions were conducted to predict healthcare utilization from the explanatory variables. Explanatory variables include age, sex, whether the provider of the biopsy is a CoC-accredited organization, Medicare dual eligibility, and the number of comorbidities. Separate models were created for the non–lung cancer patients and patients with lung cancer. Healthcare utilization was measured as the number of IP stays, OP visits, and ED visits per month of enrollment. A second aim of our study was to identify factors that influence all-cause mortality in this high-risk population. We considered individuals with a cancer diagnosis separate from those individuals without a cancer diagnosis, as it is feasible that mortality predictors in these two groups may be different. Several factors were considered for inclusion in these two models: biopsy provider (CoC organization and/or affiliated providers or non-CoC organization and/or affiliated provider), age (18–53 vs. 54–64; 53 was the median age in the dataset for all subjects), sex (male, female), IP use per month (low, high), and the number of comorbidities (1–2 vs. 3–8). IP use was defined independently for individuals with a cancer diagnosis and for those with a noncancer diagnosis. A survival model was created in SPSS software, using the COXREG procedure. All covariates shown were manually entered into the regression model in a single step. RESULTS Sample Individuals ≥ 18 to ≤ 64 years of age as of study entrance in 2013 were included. This 2013 Medicaid cohort contained 262 individuals with a lung biopsy with healthcare utilization across all hospitals, clinics, and systems in the eight counties in Western New York region. Based on inclusion and exclusion criteria, 119 (nonlung cancer) and 143 (lung cancer) patients were identified from the MDW (Figure 1). The setting for the initial biopsy was in a CoC-accredited or non-CoC organization. Fig 1 View largeDownload slide CONSORT flow diagram showing participant eligibility and assignment to lung cancer diagnosis or noncancer diagnosis group in 2013. Fig 1 View largeDownload slide CONSORT flow diagram showing participant eligibility and assignment to lung cancer diagnosis or noncancer diagnosis group in 2013. Healthcare utilization Results of the descriptive analysis for eight variables as well as the comparisons between those variables can be found in Table 2. For categorical or binary variables, gender, CoC and MCARE, the cell values in the cancer and noncancer columns are the number of events and corresponding percentage (e.g., in cancer group, 79 observations are female, which is 55.2% of the people in the cancer group). For numeric variables, the cell values in the cancer and noncancer columns are the mean and SD of the variables (e.g., in noncancer group, the patients’ mean age is 45.9, with SD = 11.5). Age, CoC accreditation affiliation, and IP utilization rate are significantly different between patients in the cancer and noncancer groups, while for other variables there is no significant difference suggested. Table 2 Two group comparisons in cancer and noncancer groups   Cancer (n = 143) n (%) or M (SD)  Noncancer (n = 119) n (%) or M (SD)  t test/chi-square statistics  p Value  Age (years)  54.7 (7.2)  45.9 (11.5)  7.23  <.001  Gender (F)  79 (55.2%)  69 (58.0%)  0.10  .749  CoC  85 (59.4%)  49 (41.2%)  7.96  .005  MCARE  57 (39.9%)  34 (28.8%)  3.17  .075  Comorbidity  2.42 (1.8)  2.39 (1.9)  0.11  .913  IP use  0.15 (0.43)  0.06 (0.09)  2.19  .030  OP use  0.93 (0.61)  0.95 (0.73)  –0.21  .831  ED use  0.16 (0.24)  0.21 (0.30)  –1.57  .119    Cancer (n = 143) n (%) or M (SD)  Noncancer (n = 119) n (%) or M (SD)  t test/chi-square statistics  p Value  Age (years)  54.7 (7.2)  45.9 (11.5)  7.23  <.001  Gender (F)  79 (55.2%)  69 (58.0%)  0.10  .749  CoC  85 (59.4%)  49 (41.2%)  7.96  .005  MCARE  57 (39.9%)  34 (28.8%)  3.17  .075  Comorbidity  2.42 (1.8)  2.39 (1.9)  0.11  .913  IP use  0.15 (0.43)  0.06 (0.09)  2.19  .030  OP use  0.93 (0.61)  0.95 (0.73)  –0.21  .831  ED use  0.16 (0.24)  0.21 (0.30)  –1.57  .119  IP inpatient; OP outpatient; ED emergency department; CoC Commission on Cancer accreditation affiliation; MCARE Medicare dual eligibility. View Large Table 2 Two group comparisons in cancer and noncancer groups   Cancer (n = 143) n (%) or M (SD)  Noncancer (n = 119) n (%) or M (SD)  t test/chi-square statistics  p Value  Age (years)  54.7 (7.2)  45.9 (11.5)  7.23  <.001  Gender (F)  79 (55.2%)  69 (58.0%)  0.10  .749  CoC  85 (59.4%)  49 (41.2%)  7.96  .005  MCARE  57 (39.9%)  34 (28.8%)  3.17  .075  Comorbidity  2.42 (1.8)  2.39 (1.9)  0.11  .913  IP use  0.15 (0.43)  0.06 (0.09)  2.19  .030  OP use  0.93 (0.61)  0.95 (0.73)  –0.21  .831  ED use  0.16 (0.24)  0.21 (0.30)  –1.57  .119    Cancer (n = 143) n (%) or M (SD)  Noncancer (n = 119) n (%) or M (SD)  t test/chi-square statistics  p Value  Age (years)  54.7 (7.2)  45.9 (11.5)  7.23  <.001  Gender (F)  79 (55.2%)  69 (58.0%)  0.10  .749  CoC  85 (59.4%)  49 (41.2%)  7.96  .005  MCARE  57 (39.9%)  34 (28.8%)  3.17  .075  Comorbidity  2.42 (1.8)  2.39 (1.9)  0.11  .913  IP use  0.15 (0.43)  0.06 (0.09)  2.19  .030  OP use  0.93 (0.61)  0.95 (0.73)  –0.21  .831  ED use  0.16 (0.24)  0.21 (0.30)  –1.57  .119  IP inpatient; OP outpatient; ED emergency department; CoC Commission on Cancer accreditation affiliation; MCARE Medicare dual eligibility. View Large Results of six regression analyses are summarized in Table 3. For patients with lung cancer, age and the number of comorbidities significantly predicted OP use and the number of comorbidities significantly predicted ED use. For non–lung cancer patients, only the number of comorbidities was the significant predictor of IP and ED uses. Explanatory variables include age, sex, whether the provider of the biopsy is a CoC-accredited organization, Medicare dual eligibility, and the number of comorbidities. Table 3 Regression model results   Cancer  Noncancer  Beta  SE  R2 change  p Value  Beta  SE  R2 change  p Value  IP use   Age  0.044  0.005  –.005  .400  –0.001  <0.001  –.016  .144   Sex (F)  –0.036  0.074  –.002  .628  –0.015  0.015  –.007  .311   CoC  0.039  0.074  –.002  .600  –0.026  0.016  –.020  .097   MCARE  –0.150  0.076  –.028  .051  –0.016  0.017  –.007  .336   Comorbidity  <0.001  0.021  <.001  .998  0.018  0.004  –.123  <.001  OP use   Age  –0.015  0.007  –.030  .028  0.010  0.006  –.022  .111   Sex (F)  0.032  0.097  –.001  .746  0.210  0.138  –.020  .130   CoC  –0.184  0.098  –.022  .063  0.044  0.143  –.001  .756   MCARE  –0.101  0.101  –.006  .316  –0.158  0.152  –.094  .303   Comorbidity  0.113  0.028  –.102  <.001  0.024  0.040  –.003  .547  ED use   Age  –0.003  0.003  –.008  .282  –0.005  0.003  –.027  .068   Sex (F)  –0.008  0.040  –.000  .841  0.038  0.055  –.004  .492   COC  –0.015  0.040  –.001  .710  –0.070  0.057  –.012  .220   MCARE  –0.042  0.042  –.007  .308  0.016  0.060  –.001  .787   Comorbidity  0.037  0.011  –.069  .002  –0.044  0.016  –.062  .006    Cancer  Noncancer  Beta  SE  R2 change  p Value  Beta  SE  R2 change  p Value  IP use   Age  0.044  0.005  –.005  .400  –0.001  <0.001  –.016  .144   Sex (F)  –0.036  0.074  –.002  .628  –0.015  0.015  –.007  .311   CoC  0.039  0.074  –.002  .600  –0.026  0.016  –.020  .097   MCARE  –0.150  0.076  –.028  .051  –0.016  0.017  –.007  .336   Comorbidity  <0.001  0.021  <.001  .998  0.018  0.004  –.123  <.001  OP use   Age  –0.015  0.007  –.030  .028  0.010  0.006  –.022  .111   Sex (F)  0.032  0.097  –.001  .746  0.210  0.138  –.020  .130   CoC  –0.184  0.098  –.022  .063  0.044  0.143  –.001  .756   MCARE  –0.101  0.101  –.006  .316  –0.158  0.152  –.094  .303   Comorbidity  0.113  0.028  –.102  <.001  0.024  0.040  –.003  .547  ED use   Age  –0.003  0.003  –.008  .282  –0.005  0.003  –.027  .068   Sex (F)  –0.008  0.040  –.000  .841  0.038  0.055  –.004  .492   COC  –0.015  0.040  –.001  .710  –0.070  0.057  –.012  .220   MCARE  –0.042  0.042  –.007  .308  0.016  0.060  –.001  .787   Comorbidity  0.037  0.011  –.069  .002  –0.044  0.016  –.062  .006  SE standard error; IP inpatient; OP outpatient; ED emergency department; CoC Commission on Cancer accreditation affiliation; MCARE Medicare dual eligibility. View Large Table 3 Regression model results   Cancer  Noncancer  Beta  SE  R2 change  p Value  Beta  SE  R2 change  p Value  IP use   Age  0.044  0.005  –.005  .400  –0.001  <0.001  –.016  .144   Sex (F)  –0.036  0.074  –.002  .628  –0.015  0.015  –.007  .311   CoC  0.039  0.074  –.002  .600  –0.026  0.016  –.020  .097   MCARE  –0.150  0.076  –.028  .051  –0.016  0.017  –.007  .336   Comorbidity  <0.001  0.021  <.001  .998  0.018  0.004  –.123  <.001  OP use   Age  –0.015  0.007  –.030  .028  0.010  0.006  –.022  .111   Sex (F)  0.032  0.097  –.001  .746  0.210  0.138  –.020  .130   CoC  –0.184  0.098  –.022  .063  0.044  0.143  –.001  .756   MCARE  –0.101  0.101  –.006  .316  –0.158  0.152  –.094  .303   Comorbidity  0.113  0.028  –.102  <.001  0.024  0.040  –.003  .547  ED use   Age  –0.003  0.003  –.008  .282  –0.005  0.003  –.027  .068   Sex (F)  –0.008  0.040  –.000  .841  0.038  0.055  –.004  .492   COC  –0.015  0.040  –.001  .710  –0.070  0.057  –.012  .220   MCARE  –0.042  0.042  –.007  .308  0.016  0.060  –.001  .787   Comorbidity  0.037  0.011  –.069  .002  –0.044  0.016  –.062  .006    Cancer  Noncancer  Beta  SE  R2 change  p Value  Beta  SE  R2 change  p Value  IP use   Age  0.044  0.005  –.005  .400  –0.001  <0.001  –.016  .144   Sex (F)  –0.036  0.074  –.002  .628  –0.015  0.015  –.007  .311   CoC  0.039  0.074  –.002  .600  –0.026  0.016  –.020  .097   MCARE  –0.150  0.076  –.028  .051  –0.016  0.017  –.007  .336   Comorbidity  <0.001  0.021  <.001  .998  0.018  0.004  –.123  <.001  OP use   Age  –0.015  0.007  –.030  .028  0.010  0.006  –.022  .111   Sex (F)  0.032  0.097  –.001  .746  0.210  0.138  –.020  .130   CoC  –0.184  0.098  –.022  .063  0.044  0.143  –.001  .756   MCARE  –0.101  0.101  –.006  .316  –0.158  0.152  –.094  .303   Comorbidity  0.113  0.028  –.102  <.001  0.024  0.040  –.003  .547  ED use   Age  –0.003  0.003  –.008  .282  –0.005  0.003  –.027  .068   Sex (F)  –0.008  0.040  –.000  .841  0.038  0.055  –.004  .492   COC  –0.015  0.040  –.001  .710  –0.070  0.057  –.012  .220   MCARE  –0.042  0.042  –.007  .308  0.016  0.060  –.001  .787   Comorbidity  0.037  0.011  –.069  .002  –0.044  0.016  –.062  .006  SE standard error; IP inpatient; OP outpatient; ED emergency department; CoC Commission on Cancer accreditation affiliation; MCARE Medicare dual eligibility. View Large Mortality risk Cox proportional hazard regression models (survival models) were developed for individuals with a lung cancer diagnosis (n = 143) and individuals without a lung cancer diagnosis (n = 119). For patients with lung cancer, significant factors include IP use per month and the number of comorbidities. Specifically, as shown in Table 4, individuals in the high IP use per month group (>0.06 encounters per month) are 2.5 times more likely to die (p < .001) than individuals in the low IP use per month group (≤0.06 encounters per month). However, individuals with a greater number of comorbid conditions (>2) were less likely to die (1–0.524 or 47.6%) (p = .005) than individuals with fewer comorbidities (1–2). Table 4 Cox proportional hazard model of time to death for individuals with a cancer diagnosis (n = 143) Factor  B  SE  Exp(B)  95% CI  p Value  Biopsy provider   Non-CoC provider (referent)      1       CoC provider  –0.23  0.22  0.79  [0.51–1.23]  .30  Age group   18–53 (referent)      1       54–64  0.29  0.23  1.34  [0.85–2.12]  .21  Sex   Male      1       Female  –0.22  0.22  0.80  [0.52–1.23]  .31  Inpatient (IP) utilization per month   Low (0.06 per month)      1       High (≥0.06 per month)  0.92  0.25  2.51  [1.55–4.07]  <.001  Number of comorbidities   Low (≤2)      1       High (>2)  –0.65  0.23  0.52  [0.33–0.83]  .005  Factor  B  SE  Exp(B)  95% CI  p Value  Biopsy provider   Non-CoC provider (referent)      1       CoC provider  –0.23  0.22  0.79  [0.51–1.23]  .30  Age group   18–53 (referent)      1       54–64  0.29  0.23  1.34  [0.85–2.12]  .21  Sex   Male      1       Female  –0.22  0.22  0.80  [0.52–1.23]  .31  Inpatient (IP) utilization per month   Low (0.06 per month)      1       High (≥0.06 per month)  0.92  0.25  2.51  [1.55–4.07]  <.001  Number of comorbidities   Low (≤2)      1       High (>2)  –0.65  0.23  0.52  [0.33–0.83]  .005  CoC Commission on Cancer, SE standard error, Exp(B) hazard ratio, CI confidence interval, degrees of freedom (df) = 1. View Large Table 4 Cox proportional hazard model of time to death for individuals with a cancer diagnosis (n = 143) Factor  B  SE  Exp(B)  95% CI  p Value  Biopsy provider   Non-CoC provider (referent)      1       CoC provider  –0.23  0.22  0.79  [0.51–1.23]  .30  Age group   18–53 (referent)      1       54–64  0.29  0.23  1.34  [0.85–2.12]  .21  Sex   Male      1       Female  –0.22  0.22  0.80  [0.52–1.23]  .31  Inpatient (IP) utilization per month   Low (0.06 per month)      1       High (≥0.06 per month)  0.92  0.25  2.51  [1.55–4.07]  <.001  Number of comorbidities   Low (≤2)      1       High (>2)  –0.65  0.23  0.52  [0.33–0.83]  .005  Factor  B  SE  Exp(B)  95% CI  p Value  Biopsy provider   Non-CoC provider (referent)      1       CoC provider  –0.23  0.22  0.79  [0.51–1.23]  .30  Age group   18–53 (referent)      1       54–64  0.29  0.23  1.34  [0.85–2.12]  .21  Sex   Male      1       Female  –0.22  0.22  0.80  [0.52–1.23]  .31  Inpatient (IP) utilization per month   Low (0.06 per month)      1       High (≥0.06 per month)  0.92  0.25  2.51  [1.55–4.07]  <.001  Number of comorbidities   Low (≤2)      1       High (>2)  –0.65  0.23  0.52  [0.33–0.83]  .005  CoC Commission on Cancer, SE standard error, Exp(B) hazard ratio, CI confidence interval, degrees of freedom (df) = 1. View Large Further, Figure 2 demonstrates the difference in hazard rate for individuals with a lung cancer diagnosis and biopsy performed by a CoC organization or affiliated provider compared with those having a biopsy performed by a non-CoC organization or affiliated provider. After 2 years, individuals with a biopsy performed by a CoC provider had approximately a 65% risk of death while those having had a biopsy by a non-CoC provider had approximately an 85% risk of death. Fig 2 View largeDownload slide Hazard function for individuals with a positive lung cancer biopsy. Fig 2 View largeDownload slide Hazard function for individuals with a positive lung cancer biopsy. The second Cox proportional hazards model included the same factors but considered only individuals without a lung cancer diagnosis. No factors were significant in this model. DISCUSSION Our findings must be placed in the context of this feasibility study to examine predictors of healthcare utilization and mortality for Medicaid individuals with or without a diagnosis of lung cancer. Our Medicaid dataset was limited to lung biopsy cases for 1 year and within a specific geographic (rural and urban) area of NY. Health utilization knowledge in low-income lung biopsy patients can provide an opportunity for providers and policy makers to improve care coordination for patients with lung cancer. Coordination of care can facilitate customized care interventions to address the specific needs of this population including surveillance following a diagnosis and SCPs for patients with lung cancer. The number and severity of comorbidities that coexist with a cancer diagnosis have been known to influence overall survival post-cancer diagnosis [22]. Our findings indicate that both the number of comorbidities and age are significant predictors of different types of healthcare services for patients with lung cancer, and only the number of comorbidities is a significant predictor of IP and ED use in non–lung cancer patients. Our findings report a higher utilization of ED and IP services among those with multiple comorbidities. Prognosis in lung cancer is determined by a myriad of factors in addition to the number of comorbidities [23]. Prior studies have proposed that mortality rates in metastatic cancer are a result of different factors including stage of disease (e.g., early vs. late), specific cancer biology, disparities in care, access to disease, or an early diagnosis in non-White patients [23, 24]. Our findings note that comorbidity influenced outcomes; however, we were not able to determine other factors because Medicaid claims data do not include stage of disease, race/ethnicity, and cancer-related biology information. In addition, we were not able to conclude whether socioeconomic status and insurance plans were contributing factors to health outcomes, as we did not compare different insurance plans. Our findings also noted that individuals with a greater number of comorbid conditions (two or more) were less likely to die than were individuals with fewer comorbidities (one to two). Our findings were not consistent with the literature on comorbidities and survival outcomes for patients with cancer [25]. The impact of all-cause survival and cancer-specific survival is likely to vary based on the type of cancer, stage of disease and prognosis, type of treatment and related sequela, and the number and severity of comorbidities [25]. Research is needed in the area of complex comorbidities in cancer and coordination of care to give clinicians opportunities for care coordination to reduce disease burden, health service utilization, and mortality rates. Our study outcome indicates that some factors related to healthcare utilization may be associated with receiving care at a CoC organization; however, with limited data we could not determine all factors related to healthcare utilization. Therefore, the specific role of the CoC accreditation status or non-CoC could not be determined as a predictor for outcomes. Fragmented survivorship care delivery is an important issue requiring new approaches to coordinated evidence-based care among those diagnosed with lung cancer [26]. As we report in our findings, patients with lung cancer used varied IP, OP, and ED healthcare services throughout the course of their illnesses. Stage of disease and symptom burden could influence the type and frequency of health service. For this study, we did not specifically track the type of each service encounter. The impact of healthcare service use on QOL was not measured in this study. Other studies have found, however, that care coordination to manage symptom burden and comorbidities is essential to improve QOL for patients with lung cancer [11]. Although our study did not stratify by stage of disease due to lack of availability in the Medicaid data or examine provider type of health service visits, it was noted that those patients with an association with a CoC organization at the time of biopsy had a greater chance of living longer than those patients associated with a non-CoC organization. Health-related issues influencing mortality in the noncancer population could be different from the health-related issues influencing mortality in the population of individuals who have a cancer diagnosis. Further, we have not yet identified in the data the reasons why Medicaid patients received a biopsy by a CoC versus a non-CoC organization. Our findings suggest that more research is needed to evaluate current quality standards such as the CoC to identify gaps and inform strategies to improve care coordination and healthcare quality. Important issues for future study will be comparisons across insurance types including employee-sponsored and publicly funded care plans as well as those individuals that would be less able to access surveillance following a diagnosis of cancer and cancer survivorship care [13, 14]. Many studies have reported disparities among patients with lung cancer including contributing factors such as race, unequal access to care, receipt of treatment, patient perception and trust, insurance status, and diagnosis at a later stage of disease [5, 8, 23, 27, 28]. Race/ethnicity, gender, socioeconomic status, and geography are significant contributors to lung cancer incidence [29]. Cancer incidence and mortality were reported higher among Black males and people of lower socioeconomic status [29, 30]. However, our Medicaid claims database did not have data on race/ethnicity. Linking cancer registry data to Medicaid claims data would address this gap in information. In addition, the average age for a lung cancer diagnosis is 65 years and older [31]. The median age for this study was 53 years. Few studies evaluate lung cancer in Medicaid patients, which insures younger and low income, relative to those in Medicare [32, 33]. Poor outcomes related to patients with lung cancer on Medicaid warrant future research to further evaluate predictors of health utilization and survival in this disparate population. Overcoming health disparities is key to improving care coordination for Medicaid patients with lung cancer. Research is needed to support ways to improve cancer detection, treatment, surveillance, QOL, and end of life care [11, 32, 33]. Limitations Several limitations for this study are noted. Although analysis of claims data has the benefit of including utilization across all settings for the vulnerable Medicaid population over multiple years, there are limitations. In the current analysis, the initial biopsy may not be the first lung biopsy and individuals may have had pre-existing lung cancer. When using a Medicaid database, it is not possible to determine the stage of cancer directly. Evidence-based guidelines (e.g., National Comprehensive Cancer Network) address all aspects of cancer disease management including screening, diagnosis, evaluation, staging, treatment, surveillance, and therapy [34]. This major limitation within the Medicaid data does not allow reliable comparisons between the treatments of patients with lung cancer and evidence-based guidelines. An overarching limitation of using a Medicaid database is the reliability for identifying cancer cases [35, 36]. Variability in cancer case identification is caused by several challenges including incomplete claims history caused by lack of continuous enrollment, age, missing procedural codes, and dual eligibility for Medicare and Medicaid programs [35, 36].In addition, challenges exist for the predictive ability of comorbidity and diagnostic-based indices, as well as how sensitive and specific these are in predicting outcomes related to comorbidity, mortality, and health utilization for patients with cancer [37]. Predictive ability of comorbidity indices can vary widely depending on the specific index (e.g., Charlson Comorbidity Index and Elixhauser Index), and selection of the appropriate index for use should consider type of available data, study population, and specific outcome(s) for the study [37]. Combining Medicaid data sources and the cancer registry data would provide a more complete and robust data, accounting for variables found exclusively in one or the other data set (e.g., stage of disease, race/ethnicity, socioeconomic status) [38]. Medicaid data provide coverage across populations when access to other databases such as SEER or a cancer registry database is not available [39, 40]. Therefore, linking to cancer registry data may be better option for combining data sources. Future studies linking cancer registry databases with Medicaid data would provide additional patient information including diagnosis, stage of disease, date of diagnosis, and cause of mortality providing opportunities to assess cancer care [41]. The average age for a lung cancer diagnosis is 65 years or older [31]; therefore, age may be considered a limitation for this study due to relatively fewer number of lung cancer cases in patients ≥18 to ≤64 years of age. Medicaid claims data can provide an opportunity to evaluate health service utilization and outcomes for future research directed toward a low-income and younger lung cancer population that may benefit from new treatments, health behavior change (e.g., tobacco cessation programs), and supportive care to improve QOL. The time for the calculation for mortality is short, and future research would benefit from a larger sample size and extended timeline for surveillance of data. The analysis is limited to the eight counties of Western New York where we have knowledge of affiliation with cancer centers. CONCLUSION Care coordination is essential for low-income individuals at risk or with a diagnosis of lung cancer. Patients in this population burdened with symptoms of their disease that affect overall QOL could benefit from appropriate coordination of care. Evidence-based lung cancer guidelines have been created to provide the most up-to-date information intended to improve the management of patient care as well as patient outcomes [2–9]. A recent series of articles published on cancer survivorship in the USA discussed how cancer survivorship is defined, the ongoing needs and preferences over the continuum of survivorship care, and considerations for coordinated cancer care [42–44]. To our knowledge, few studies have examined health service utilization of patients with lung cancer to explore opportunities for chronic care coordination based on type and frequency of service. Future research is needed to explore how patterns of health service utilization can be useful in developing and coordinating care plans across the cancer survivor continuum. Primary Data: Findings reported in this manuscript have not been previously published, and the manuscript is not being simultaneously submitted elsewhere. There has been no previous reporting of data; the authors have full control of all primary data, and the authors of this manuscript agree to allow the journal to review their data if requested. The authors have no study funding sources to report. Compliance with Ethical Standards Conflict of Interest: Authors Somayaji, Chang, Casucci, Xue, and Hewner declare that they have no conflict of interest. Ethical Approval: This article does not contain any studies with human participants or animals performed by any of the authors. Based on the data extraction and de-identification protocol, the University at Buffalo Institutional Review Board determined that this project did not qualify as human subjects research. Informed Consent: For this type of study, formal consent is not required. References 1. Williams DR, Kontos EZ, Viswanath Ket al.   Integrating multiple social statuses in health disparities research: The case of lung cancer. Health Serv Res . 2012; 47( 3 pt 2): 1255– 1277. Google Scholar CrossRef Search ADS PubMed  2. Alberg AJ, Brock MV, Ford JG, Samet JM, Spivack SD. Epidemiology of lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest . 2013; 143( 5 suppl): e1S– e29S. Google Scholar CrossRef Search ADS PubMed  3. McDonald KM, Sundaram V, Bravata DM,et al.   Closing the quality gap: A critical analysis of quality improvement strategies (vol. 7: care coordination). Rockville, MD: Agency for Healthcare Research and Quality (US); 2007 Jun. (Technical Reviews, No. 9.7.) Available from:https://www.ncbi.nlm.nih.gov/books/NBK44015/ 4. Rudin, Robert S. Courtney A. Gidengil, Zachary Predmore, Eric C. Schneider, James Sorace, and Rachel Hornstein, Identifying and Coordinating Care for Complex Patients: Findings from the Leading Edge of Analytics and Health Information Technology. Santa Monica, CA: RAND Corporation, 2016. Available from:https://www.rand.org/pubs/research_reports/RR1234.html 5. Slatore CG, Au DH, Gould MK; American Thoracic Society Disparities in Healthcare Group. An official American Thoracic Society systematic review: Insurance status and disparities in lung cancer practices and outcomes. Am J Respir Crit Care Med . 2010; 182( 9): 1195– 1205. Google Scholar CrossRef Search ADS PubMed  6. Mayer DK, Gerstel A, Walton ALet al.   Implementing survivorship care plans for colon cancer survivors. Oncol Nurs Forum . 2014; 41( 3): 266– 273. Google Scholar CrossRef Search ADS PubMed  7. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin . 2016; 66( 1): 7– 30. Google Scholar CrossRef Search ADS PubMed  8. Abdelsattar ZM, Hendren S, Wong SL. The impact of health insurance on cancer care in disadvantaged communities. Cancer . 2017; 123( 7): 1219– 1227. Google Scholar CrossRef Search ADS PubMed  9. Detterbeck FC, Lewis SZ, Diekemper R, Addrizzo-Harris D, Alberts WM. Executive summary: Diagnosis and management of lung cancer: American College of Chest Physicians evidence-based clinical practice guidelines. Chest . 2013; 143( 5 suppl): 7S– 37S. Google Scholar CrossRef Search ADS PubMed  10. Kaniski F, Enewold L, Thomas A, Malik S, Stevens JL, Harlan LC. Temporal patterns of care and outcomes of non-small cell lung cancer patients in the United States diagnosed in 1996, 2005, and 2010. Lung cancer . 2017; 103( suppl C): 66– 74. Google Scholar CrossRef Search ADS PubMed  11. Yang P, Cheville AL, Wampfler JAet al.   Quality of life and symptom burden among long-term lung cancer survivors. J Thorac Oncol . 2012; 7( 1): 64– 70. Google Scholar CrossRef Search ADS PubMed  12. Iyer S, Roughley A, Rider A, Taylor-Stokes G. The symptom burden of non-small cell lung cancer in the USA: A real-world cross-sectional study. Support Care Cancer . 2014; 22( 1): 181– 187. Google Scholar CrossRef Search ADS PubMed  13. Kehl KL, Liao K-P, Krause TM, Giordano SH. Access to accredited cancer hospitals within federal exchange plans under the Affordable Care Act. J Clin Oncol . 2017; 35( 6): 645– 651. Google Scholar CrossRef Search ADS PubMed  14. Graves JA, Swartz K. Effects of Affordable Care Act marketplaces and Medicaid eligibility expansion on access to cancer care. Cancer J . 2017; 23( 3): 168– 174. Google Scholar CrossRef Search ADS PubMed  15. Oncology ASoC. The state of cancer care in America, 2015: A report by the American Society of Clinical Oncology. J Oncol Pract . 2015; 11( 2): 79– 113. CrossRef Search ADS PubMed  16. Clarke JL, Bourn S, Skoufalos A, Beck EH, Castillo DJ. An innovative approach to health care delivery for patients with chronic conditions. Popul Health Manag . 2017; 20( 1): 23– 30. Google Scholar CrossRef Search ADS PubMed  17. McCabe MS, Bhatia S, Oeffinger KCet al.   American society of clinical oncology statement: Achieving high-quality cancer survivorship care. J Clin Oncol . 2013; 31( 5): 631– 640. Google Scholar CrossRef Search ADS PubMed  18. David EA, Cooke DT, Chen Y, Perry A, Canter RJ, Cress R. Surgery in high-volume hospitals not Commission on Cancer accreditation leads to increased cancer-specific survival for early-stage lung cancer. Am J Surg . 2015; 210( 4): 643– 647. Google Scholar CrossRef Search ADS PubMed  19. Cancer ACoSCo. Cancer Program Standards: Ensuring Patient-Centered Care , American College of Surgeons Commission on Cancer, Cancer Program Standards: Ensuring Patient-Centered Care, 2016 ed. 20. Bilimoria KY, Bentrem DJ, Stewart AK, Winchester DP, Ko CY. Comparison of commission on cancer-approved and -nonapproved hospitals in the United States: Implications for studies that use the National Cancer Data Base. J Clin Oncol . 2009; 27( 25): 4177– 4181. Google Scholar CrossRef Search ADS PubMed  21. H-CUP. Clinical Classifications Software for ICD-9-CM. 2016; Available at https://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp. Accessibility verified January 31, 2017. 22. Edwards BK, Noone AM, Mariotto ABet al.   Annual report to the nation on the status of cancer, 1975–2010, featuring prevalence of comorbidity and impact on survival among persons with lung, colorectal, breast, or prostate cancer. Cancer . 2014; 120( 9): 1290– 1314. Google Scholar CrossRef Search ADS PubMed  23. Grose D, Morrison DS, Devereux Get al.  ; Scottish Lung Cancer Forum. The impact of comorbidity upon determinants of outcome in patients with lung cancer. Lung Cancer . 2015; 87( 2): 186– 192. Google Scholar CrossRef Search ADS PubMed  24. Loh KP, Kansagra A, Shieh MSet al.   Predictors of in-hospital mortality in patients with metastatic cancer receiving specific critical care therapies. J Natl Compr Canc Netw . 2016; 14( 8): 979– 987. Google Scholar CrossRef Search ADS PubMed  25. Sarfati D, Koczwara B, Jackson C. The impact of comorbidity on cancer and its treatment. CA Cancer J Clin . 2016; 66( 4): 337– 350. Google Scholar CrossRef Search ADS PubMed  26. Klabunde CN, Haggstrom D, Kahn KL, Gray SW, Kim B, Liu B, Eisenstein J, Keating NL. Oncologists’ perspectives on post‐cancer treatment communication and care coordination with primary care physicians. Eur J of Cancer Care. 2017;26(4). 27. Yang R, Cheung MC, Byrne MMet al.   Do racial or socioeconomic disparities exist in lung cancer treatment? Cancer . 2010; 116( 10): 2437– 2447. Google Scholar PubMed  28. Niu X, Roche LM, Pawlish KS, Henry KA. Cancer survival disparities by health insurance status. Cancer Med . 2013; 2( 3): 403– 411. Google Scholar CrossRef Search ADS PubMed  29. Tabatabai MA, Kengwoung-Keumo JJ, Oates GRet al.   Racial and gender disparities in incidence of lung and bronchus cancer in the united states: A longitudinal analysis. Plos One . 2016; 11( 9): e0162949. Google Scholar CrossRef Search ADS PubMed  30. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2017. CA Cancer J Clin . 2017; 67( 1): 7– 30. Google Scholar CrossRef Search ADS PubMed  31. Society AC. Key Statistics for Lung Cancer. 2017; Available at https://www.cancer.org/cancer/non-small-cell-lung-cancer/about/key-statistics.html. Accessibility verified December 20, 2017. 32. Mack JW, Chen K, Boscoe FPet al.   Underuse of hospice care by Medicaid-insured patients with stage IV lung cancer in New York and California. J Clin Oncol . 2013; 31( 20): 2569– 2579. Google Scholar CrossRef Search ADS PubMed  33. Walker GV, Grant SR, Guadagnolo BAet al.   Disparities in stage at diagnosis, treatment, and survival in nonelderly adult patients with cancer according to insurance status. J Clin Oncol . 2014; 32( 28): 3118– 3125. Google Scholar CrossRef Search ADS PubMed  34. National Comprehensive Cancer Network. National Comprehensive Network Guidelines. 2018; Available at https://www.nccn.org/. Accessibility verified February 7, 2018. 35. Koroukian SM, Cooper GS, Rimm AA. Ability of medicaid claims data to identify incident cases of breast cancer in the Ohio Medicaid population. Health Serv Res . 2003; 38( 3): 947– 960. Google Scholar CrossRef Search ADS PubMed  36. Schrag D, Virnig BA, Warren JL. Linking tumor registry and medicaid claims to evaluate cancer care delivery. Health Care Financ Rev . 2009; 30( 4): 61– 73. Google Scholar PubMed  37. Yurkovich M, Avina-Zubieta JA, Thomas J, Gorenchtein M, Lacaille D. A systematic review identifies valid comorbidity indices derived from administrative health data. J Clin Epidemiol . 2015; 68( 1): 3– 14. Google Scholar CrossRef Search ADS PubMed  38. Schrag D, Virnig BA, Warren JL. Linking tumor registry and medicaid claims to evaluate cancer care delivery. Health Care Financ Rev . 2009; 30( 4): 61– 73. Google Scholar PubMed  39. Institute NC. List of SEER Registries. 2017; Available at https://seer.cancer.gov/registries/list.html. Accessibility verified December 22, 2017. 40. Sciences NCIDoCCP. National and State Cancer Registries. 2017; Available at https://epi.grants.cancer.gov/registries.html. Accessibility verified December 22, 2017. 41. Boscoe FP, Schrag D, Chen K, Roohan PJ, Schymura MJ. Building capacity to assess cancer care in the medicaid population in New York State. Health Serv Res . 2011; 46( 3): 805– 820. Google Scholar CrossRef Search ADS PubMed  42. Jacobs LA, Shulman LN. Follow-up care of cancer survivors: Challenges and solutions. Lancet Oncol . 2017; 18( 1): e19– e29. Google Scholar CrossRef Search ADS PubMed  43. Mayer DK, Nasso SF, Earp JA. Defining cancer survivors, their needs, and perspectives on survivorship health care in the USA. Lancet Oncol . 2017; 18( 1): e11– e18. Google Scholar CrossRef Search ADS PubMed  44. Nekhlyudov L, O’malley DM, Hudson SV. Integrating primary care providers in the care of cancer survivors: Gaps in evidence and future opportunities. Lancet Oncol . 2017; 18( 1): e30– e38. Google Scholar CrossRef Search ADS PubMed  © Society of Behavioral Medicine 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Translational Behavioral Medicine Oxford University Press

Exploring Medicaid claims data to understand predictors of healthcare utilization and mortality for Medicaid individuals with or without a diagnosis of lung cancer: a feasibility study

Loading next page...
 
/lp/ou_press/exploring-medicaid-claims-data-to-understand-predictors-of-healthcare-PeyaPhossw
Copyright
© Society of Behavioral Medicine 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
1869-6716
eISSN
1613-9860
D.O.I.
10.1093/tbm/iby023
Publisher site
See Article on Publisher Site

Abstract

Abstract Health disparities in low-income populations complicate care for at-risk individuals or those diagnosed with lung cancer and may influence their patterns of healthcare utilization. The purpose of this study is to examine whether age, sex, provider’s affiliation, Medicare dual eligibility, and number of comorbidities can predict healthcare utilization, as well as to examine factors influencing mortality in lung biopsy patients. A retrospective review of de-identified Medicaid claims of adults having a lung biopsy in 2013 resulted in classification into lung cancer and non–lung cancer cases based on a lung cancer diagnostic code within 30 days after biopsy. Biopsy cases were further divided by whether or not the provider’s institution was accredited by the Commission on Cancer (CoC). Inpatient (IP), outpatient (OP), and emergency department (ED) utilization was followed from initial date of biopsy through 2015, or to the earliest date of death, disenrollment, or study end for both groups. The result of Cox proportional hazards regression model indicated that age and the number of comorbidities significantly predicted OP use and the number of comorbidities significantly predicted ED use in patients with lung cancer. However, for non–lung cancer patients, only the number of comorbidities significantly predicted IP and ED uses. Furthermore, for patients with lung cancer, the significant factors of mortality included IP use per month and the number of comorbidities. Patients with lung cancer who received a lung biopsy by a CoC-accredited organization had a longer time of survival from the biopsy event. Our findings suggest that understanding predictors of healthcare utilization and mortality may create opportunities to improve health and quality of life through better healthcare coordination. Implications Research: Future research is needed to examine healthcare utilization patterns with recommended evidence-based guidelines for patients with lung cancer on Medicaid. Practice: Understanding healthcare utilization patterns for low-income patients with lung cancer can provide opportunities to identify healthcare disparities in access, cost, and quality and can further support the development of appropriate and effective interventions to improve overall quality of care. Policy: Health organizations and policy makers may need to take new approaches to promote data sharing to create robust databases to facilitate lung cancer research that improves quality of care. INTRODUCTION Limited access to health resources and poor coordination of healthcare in low-income populations is associated with underuse of cancer screening, higher risk for a late diagnosis of cancer, increased symptoms related to a cancer diagnosis and comorbid disease, decreased quality of life (QOL), and overall survival [1, 2]. Care coordination is the organization of patient care activities including access to resources and exchange of critical information to facilitate the appropriate delivery of healthcare services [3]. Several factors influence health outcomes including lack of information at point of care, as well as poor communication between primary and specialty care relating to inpatient (IP) stays, outpatient (OP) visits, and emergency department (ED) visits. Integrating coordination programs continues to be challenging at the provider level as care delivery processes and plans can be both time-consuming and financially burdensome [4, 5]. Although there have been initiatives over the past decade to develop cancer survivorship care plans (SCPs) [6], there is little known regarding healthcare utilization and care coordination for lung cancer. Lung cancer is the most common cause of cancer death among the four major cancers [7]. Cancer care disparities are complex and not well understood; some of the reasons for this are patient related including their health behavior, healthcare provider, as well as their varying capacity to interact with the healthcare system [5, 8]. Attention to lung cancer has increased with new advances and collaborative work among clinicians and scientists; however, health disparities still exist in managing health-directed treatment, comorbidities, QOL, and symptom management [5, 8–11]. Health-related QOL for lung cancer survivors are not equitable to age-matched survivors of other cancers [11]. Utilization of appropriate healthcare services may help to address such issues as compromised lung capacity, overwhelming symptom burden, and treatment-related morbidity affecting QOL [11, 12]. In addition, the burden of disease varies among populations based on insurance status [5]. Differences in groups are evident in mortality, access to care, and delivery of care. Patients with lung cancer with Medicaid insurance have higher incidence rates, poorer outcomes with more advanced disease at diagnosis, and poorer survival. Patients with Medicaid are more likely than others to die during the month they were diagnosed and are less likely to receive surgery, radiation, or treatment at centers managing a high volume of patients [5]. The changing landscape of the Affordable Care Act has created angst among many health service providers regarding access to specialized cancer care especially for Medicaid beneficiaries [13–15]. Gaps and restrictions in Medicaid coverage can create barriers facilitating continued disparities in access to care, narrowed provider networks to provide specialized cancer services, and meeting evidence-based guidelines for quality care [15]. Quality measures and national benchmarks are essential tools to ensure evidence-based guideline care for patients with cancer [15–17]. The Commission on Cancer (CoC) was established by the American College of Surgeons to provide an accredited process to evaluate oncologic outcomes and the QOL for patients with cancer and to ensure continuity of care and access to cancer resources and services [18–20]. CoC accreditation is based on compliance and adherence to recommended guidelines for cancer, establishing the benchmark for delivery of specialized care, and addressing health disparities and barriers to care which is important when evaluating health service utilization and quality monitoring of individuals enrolled in Medicaid. Low-income and younger lung cancer survivors may not utilize appropriate healthcare services and therefore may not be meeting national evidence-based recommendations for disease management of prevention, screening, treatment, supportive, and survivorship care [1, 10]. Reducing health disparities underscores the critical need for understanding health utilization and lung cancer care in a younger, diverse, low-income, understudied, and higher risk population such as those participating in a Medicaid insurance program. Purpose The purpose of this study is to examine whether age, sex, provider’s affiliation, Medicare dual eligibility, and the number of comorbidities can predict healthcare utilization, as well as to examine factors influencing mortality in lung biopsy patients. METHODS Study design The population, variables, and analysis selected address the gaps in healthcare utilization in lung biopsy cases. The analysis compares individuals with a biopsy followed by a lung cancer diagnosis to individuals with biopsy without a cancer diagnosis. We conducted a retrospective review of de-identified Medicaid claims of adults in eight counties in Western New York (Erie, Niagara, Orleans, Genesee, Wyoming, Allegany, Cattaraugus, and Chautauqua) who had a lung biopsy between January 1 and December 31, 2013. Cases were followed from initial date of biopsy through December 31, 2015 or to the earliest date of death, disenrollment, or study end for both groups. We stratified both lung cancer and noncancer cases by initial biopsy in CoC (CoC accreditation status) and non-CoC accreditation. We captured data on IP, OP, and ED visits and mortality. The University at Buffalo Institutional Review Board reviewed the protocol for de-identification and approved it as not human subject research. Data source The secondary data analysis utilized existing de-identified demographics and claims data extracted from the Medicaid Data Warehouse (MDW). The claims data include the same research ID, a flag for type of claim (e.g., IP, ED, OP), dates-of-service converted to interval from a random date, a flag for the provider of a specific service, and the first five diagnosis codes. As the MDW contains protected health information, analysts extracted Health Insurance Portability and Accountability Act of 1996 (HIPAA) compliant demographic and claims data and assigned each case a unique research ID. The demographic database consists of research ID, age, sex, county of residence, Medicare eligibility, and a flag for death within a year. Measures We used the International Classification of Disease based on the Healthcare Cost and Utilization Project (H-CUP) definition for lung cancer and major chronic conditions. H-CUP’s Clinical Classification Software (CCS) aggregates diagnosis codes into conditions [21]. Chronic conditions were divided into two groups based on the number of comorbidities (one to two comorbidities, and three to eight comorbidities). See Table 1 for details of codes included in cancer and chronic disease groups. Table 1 International Classification of Disease (ICD)-9 and Current Procedural Terminology (CPT) codes used to identify patients with cancer of bronchus and lung Domain  CCS codes for ICD-9  Descriptions  Chronic disease base  49, 50, 186 (DM); 66, 67 (SA); 65, 68, 69, 70, 71, 72, 73, 74, 75 (MH); 98, 183 (HTN); 99a, 158 (CKD); 103, 108 (HF); 100, 101 (CAD); 127 (COPD); 128 (ASTH)  Diabetes (DM); substance abuse (SA); mental health (MH); hypertension (HTN); chronic kidney disease (CKD); heart failure (HF); coronary artery disease (CAD); chronic obstructive pulmonary disease (COPD); asthma (ASTH)  Cancer  13; 14; 15; 17; 19,20; 22; 24; 25; 27; 29; 30; 33; 36  Stomach; colon; rectal; pancreatic; lung; melanoma; breast; uterine; ovarian; prostate; bladder; kidney; thyroid  Bronchus and lung cancer  1622  MALIG NEO MAIN BRONCHUS    1623  MAL NEO UPPER LOBE LUNG    1624  MAL NEO MIDDLE LOBE LUNG    1625  MAL NEO LOWER LOBE LUNG    1628  MAL NEO BRONCH/LUNG NEC    1629  MAL NEO BRONCH/LUNG NOS    20921  MAL CARCINOID BRONC/LUNG    2312  CA IN SITU BRONCHUS/LUNG    CPT codes for lung biopsy  Description  Codes for lung biopsy  31628  Bronchoscopy with transbronchial lung biopsy(ies) single lobe    31629  Bronchoscopy with transbronchial needle aspiration biopsy(ies) trachea, main stem and/or lobar bronchus(i)    31632  Bronchoscopy with transbronchial lung biopsy(ies), each additional lobe    31633  Bronchoscopy with transbronchial needle aspiration biopsy(ies), each additional lobe    32096  Thoracotomy with diagnostic biopsy(ies) of lung infiltrates (e.g., wedge, incisional), unilateral    32097  Thoracotomy with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32098  Thoracotomy with biopsy(ies) of pleura    32400  Biopsy pleura; percutaneous needle    32405  Biopsy lung or mediastinum, percutaneous needle    32606  Thoracoscopy, mediastinal space, with biopsy    32607  Thoracoscopy; with diagnostic biopsy(ies) of lung infiltrate(s) (e.g., wedge, incisional), unilateral    32608  Thoracoscopy; with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32609  Thoracoscopy; with biopsy(ies) of pleura    39000  Open biopsy of mediastinum    39010  Open biopsy of mediastinum  Domain  CCS codes for ICD-9  Descriptions  Chronic disease base  49, 50, 186 (DM); 66, 67 (SA); 65, 68, 69, 70, 71, 72, 73, 74, 75 (MH); 98, 183 (HTN); 99a, 158 (CKD); 103, 108 (HF); 100, 101 (CAD); 127 (COPD); 128 (ASTH)  Diabetes (DM); substance abuse (SA); mental health (MH); hypertension (HTN); chronic kidney disease (CKD); heart failure (HF); coronary artery disease (CAD); chronic obstructive pulmonary disease (COPD); asthma (ASTH)  Cancer  13; 14; 15; 17; 19,20; 22; 24; 25; 27; 29; 30; 33; 36  Stomach; colon; rectal; pancreatic; lung; melanoma; breast; uterine; ovarian; prostate; bladder; kidney; thyroid  Bronchus and lung cancer  1622  MALIG NEO MAIN BRONCHUS    1623  MAL NEO UPPER LOBE LUNG    1624  MAL NEO MIDDLE LOBE LUNG    1625  MAL NEO LOWER LOBE LUNG    1628  MAL NEO BRONCH/LUNG NEC    1629  MAL NEO BRONCH/LUNG NOS    20921  MAL CARCINOID BRONC/LUNG    2312  CA IN SITU BRONCHUS/LUNG    CPT codes for lung biopsy  Description  Codes for lung biopsy  31628  Bronchoscopy with transbronchial lung biopsy(ies) single lobe    31629  Bronchoscopy with transbronchial needle aspiration biopsy(ies) trachea, main stem and/or lobar bronchus(i)    31632  Bronchoscopy with transbronchial lung biopsy(ies), each additional lobe    31633  Bronchoscopy with transbronchial needle aspiration biopsy(ies), each additional lobe    32096  Thoracotomy with diagnostic biopsy(ies) of lung infiltrates (e.g., wedge, incisional), unilateral    32097  Thoracotomy with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32098  Thoracotomy with biopsy(ies) of pleura    32400  Biopsy pleura; percutaneous needle    32405  Biopsy lung or mediastinum, percutaneous needle    32606  Thoracoscopy, mediastinal space, with biopsy    32607  Thoracoscopy; with diagnostic biopsy(ies) of lung infiltrate(s) (e.g., wedge, incisional), unilateral    32608  Thoracoscopy; with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32609  Thoracoscopy; with biopsy(ies) of pleura    39000  Open biopsy of mediastinum    39010  Open biopsy of mediastinum  CCS Clinical Classification Software. aCodes for HTN, CKD, and HF. View Large Table 1 International Classification of Disease (ICD)-9 and Current Procedural Terminology (CPT) codes used to identify patients with cancer of bronchus and lung Domain  CCS codes for ICD-9  Descriptions  Chronic disease base  49, 50, 186 (DM); 66, 67 (SA); 65, 68, 69, 70, 71, 72, 73, 74, 75 (MH); 98, 183 (HTN); 99a, 158 (CKD); 103, 108 (HF); 100, 101 (CAD); 127 (COPD); 128 (ASTH)  Diabetes (DM); substance abuse (SA); mental health (MH); hypertension (HTN); chronic kidney disease (CKD); heart failure (HF); coronary artery disease (CAD); chronic obstructive pulmonary disease (COPD); asthma (ASTH)  Cancer  13; 14; 15; 17; 19,20; 22; 24; 25; 27; 29; 30; 33; 36  Stomach; colon; rectal; pancreatic; lung; melanoma; breast; uterine; ovarian; prostate; bladder; kidney; thyroid  Bronchus and lung cancer  1622  MALIG NEO MAIN BRONCHUS    1623  MAL NEO UPPER LOBE LUNG    1624  MAL NEO MIDDLE LOBE LUNG    1625  MAL NEO LOWER LOBE LUNG    1628  MAL NEO BRONCH/LUNG NEC    1629  MAL NEO BRONCH/LUNG NOS    20921  MAL CARCINOID BRONC/LUNG    2312  CA IN SITU BRONCHUS/LUNG    CPT codes for lung biopsy  Description  Codes for lung biopsy  31628  Bronchoscopy with transbronchial lung biopsy(ies) single lobe    31629  Bronchoscopy with transbronchial needle aspiration biopsy(ies) trachea, main stem and/or lobar bronchus(i)    31632  Bronchoscopy with transbronchial lung biopsy(ies), each additional lobe    31633  Bronchoscopy with transbronchial needle aspiration biopsy(ies), each additional lobe    32096  Thoracotomy with diagnostic biopsy(ies) of lung infiltrates (e.g., wedge, incisional), unilateral    32097  Thoracotomy with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32098  Thoracotomy with biopsy(ies) of pleura    32400  Biopsy pleura; percutaneous needle    32405  Biopsy lung or mediastinum, percutaneous needle    32606  Thoracoscopy, mediastinal space, with biopsy    32607  Thoracoscopy; with diagnostic biopsy(ies) of lung infiltrate(s) (e.g., wedge, incisional), unilateral    32608  Thoracoscopy; with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32609  Thoracoscopy; with biopsy(ies) of pleura    39000  Open biopsy of mediastinum    39010  Open biopsy of mediastinum  Domain  CCS codes for ICD-9  Descriptions  Chronic disease base  49, 50, 186 (DM); 66, 67 (SA); 65, 68, 69, 70, 71, 72, 73, 74, 75 (MH); 98, 183 (HTN); 99a, 158 (CKD); 103, 108 (HF); 100, 101 (CAD); 127 (COPD); 128 (ASTH)  Diabetes (DM); substance abuse (SA); mental health (MH); hypertension (HTN); chronic kidney disease (CKD); heart failure (HF); coronary artery disease (CAD); chronic obstructive pulmonary disease (COPD); asthma (ASTH)  Cancer  13; 14; 15; 17; 19,20; 22; 24; 25; 27; 29; 30; 33; 36  Stomach; colon; rectal; pancreatic; lung; melanoma; breast; uterine; ovarian; prostate; bladder; kidney; thyroid  Bronchus and lung cancer  1622  MALIG NEO MAIN BRONCHUS    1623  MAL NEO UPPER LOBE LUNG    1624  MAL NEO MIDDLE LOBE LUNG    1625  MAL NEO LOWER LOBE LUNG    1628  MAL NEO BRONCH/LUNG NEC    1629  MAL NEO BRONCH/LUNG NOS    20921  MAL CARCINOID BRONC/LUNG    2312  CA IN SITU BRONCHUS/LUNG    CPT codes for lung biopsy  Description  Codes for lung biopsy  31628  Bronchoscopy with transbronchial lung biopsy(ies) single lobe    31629  Bronchoscopy with transbronchial needle aspiration biopsy(ies) trachea, main stem and/or lobar bronchus(i)    31632  Bronchoscopy with transbronchial lung biopsy(ies), each additional lobe    31633  Bronchoscopy with transbronchial needle aspiration biopsy(ies), each additional lobe    32096  Thoracotomy with diagnostic biopsy(ies) of lung infiltrates (e.g., wedge, incisional), unilateral    32097  Thoracotomy with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32098  Thoracotomy with biopsy(ies) of pleura    32400  Biopsy pleura; percutaneous needle    32405  Biopsy lung or mediastinum, percutaneous needle    32606  Thoracoscopy, mediastinal space, with biopsy    32607  Thoracoscopy; with diagnostic biopsy(ies) of lung infiltrate(s) (e.g., wedge, incisional), unilateral    32608  Thoracoscopy; with diagnostic biopsy(ies) of lung nodule(s) or mass(es) (e.g., wedge, incisional), unilateral    32609  Thoracoscopy; with biopsy(ies) of pleura    39000  Open biopsy of mediastinum    39010  Open biopsy of mediastinum  CCS Clinical Classification Software. aCodes for HTN, CKD, and HF. View Large We identified the presence or absence of lung cancer by examining de-identified Medicaid claims of adults (≥18 to ≤64 years of age) having a lung biopsy in 2013. Specific Current Procedural Terminology (CPT) codes, as shown in Table 1, were used to identify patients admitted for a lung biopsy. As cancer stage and initial date of primary diagnosis were not available, a proxy date of lung cancer diagnosis was created using a 30-day window post diagnostic lung biopsy. This procedure created two groups: lung cancer cases and non–lung cancer cases. This approach provided a practical solution for identifying individuals with a lung cancer diagnosis and for identifying lung biopsy patients who may be a population at risk for lung cancer. This approach was particularly useful for facilitating analysis when specific cancer registries or detailed patient data are not available. In addition, as Medicaid data only include date of death, we used all-cause mortality rate as a measure of the number of deaths in this study. Utilization was based on data definitions for IP, ED, and OP events. IP events were identified for all claims with a hospital listed as the place of service and eliminating all nonadmission events, such as, emergency care, laboratory and imaging studies, and other clinic visits. ED and OP events were similarly identified using evaluation and management CPT codes specific to those settings. Utilization for each person with a biopsy was tracked from the initial date of biopsy to the earliest of death, disenrollment, or end of study for both groups. We created a flag for the provider to indicate whether or not the biopsy was completed in a CoC-accredited organization. Data analysis The unit of analysis was the individual Medicaid recipient who represented a lung biopsy case. Descriptive analyses summarized demographic characteristics. The data are presented as means and standard deviations (SD) for continuous variables, as well as frequencies and percentages for categorical variables. Two group comparisons were conducted to evaluate the overall value of each variable for a significant difference in cancer and noncancer groups. For each of the two groups (i.e., cancer and noncancer), we conducted a descriptive analysis for eight variables and obtained a mean and an SD values for numeric (continuous) variables and the number of occurrence and percentage for categorical (factor) variables. For categorical or binary variables, gender, CoC, and MCARE, the cell values in the cancer and noncancer columns are the number of events and corresponding percentage, and the chi-square test was used to make the comparison of categorical variables. For numeric variables, the cell values in the cancer and noncancer columns are the mean and SD of the variables, and we used t test to test the differences between these numeric variables. Multiple linear regressions were conducted to predict healthcare utilization from the explanatory variables. Explanatory variables include age, sex, whether the provider of the biopsy is a CoC-accredited organization, Medicare dual eligibility, and the number of comorbidities. Separate models were created for the non–lung cancer patients and patients with lung cancer. Healthcare utilization was measured as the number of IP stays, OP visits, and ED visits per month of enrollment. A second aim of our study was to identify factors that influence all-cause mortality in this high-risk population. We considered individuals with a cancer diagnosis separate from those individuals without a cancer diagnosis, as it is feasible that mortality predictors in these two groups may be different. Several factors were considered for inclusion in these two models: biopsy provider (CoC organization and/or affiliated providers or non-CoC organization and/or affiliated provider), age (18–53 vs. 54–64; 53 was the median age in the dataset for all subjects), sex (male, female), IP use per month (low, high), and the number of comorbidities (1–2 vs. 3–8). IP use was defined independently for individuals with a cancer diagnosis and for those with a noncancer diagnosis. A survival model was created in SPSS software, using the COXREG procedure. All covariates shown were manually entered into the regression model in a single step. RESULTS Sample Individuals ≥ 18 to ≤ 64 years of age as of study entrance in 2013 were included. This 2013 Medicaid cohort contained 262 individuals with a lung biopsy with healthcare utilization across all hospitals, clinics, and systems in the eight counties in Western New York region. Based on inclusion and exclusion criteria, 119 (nonlung cancer) and 143 (lung cancer) patients were identified from the MDW (Figure 1). The setting for the initial biopsy was in a CoC-accredited or non-CoC organization. Fig 1 View largeDownload slide CONSORT flow diagram showing participant eligibility and assignment to lung cancer diagnosis or noncancer diagnosis group in 2013. Fig 1 View largeDownload slide CONSORT flow diagram showing participant eligibility and assignment to lung cancer diagnosis or noncancer diagnosis group in 2013. Healthcare utilization Results of the descriptive analysis for eight variables as well as the comparisons between those variables can be found in Table 2. For categorical or binary variables, gender, CoC and MCARE, the cell values in the cancer and noncancer columns are the number of events and corresponding percentage (e.g., in cancer group, 79 observations are female, which is 55.2% of the people in the cancer group). For numeric variables, the cell values in the cancer and noncancer columns are the mean and SD of the variables (e.g., in noncancer group, the patients’ mean age is 45.9, with SD = 11.5). Age, CoC accreditation affiliation, and IP utilization rate are significantly different between patients in the cancer and noncancer groups, while for other variables there is no significant difference suggested. Table 2 Two group comparisons in cancer and noncancer groups   Cancer (n = 143) n (%) or M (SD)  Noncancer (n = 119) n (%) or M (SD)  t test/chi-square statistics  p Value  Age (years)  54.7 (7.2)  45.9 (11.5)  7.23  <.001  Gender (F)  79 (55.2%)  69 (58.0%)  0.10  .749  CoC  85 (59.4%)  49 (41.2%)  7.96  .005  MCARE  57 (39.9%)  34 (28.8%)  3.17  .075  Comorbidity  2.42 (1.8)  2.39 (1.9)  0.11  .913  IP use  0.15 (0.43)  0.06 (0.09)  2.19  .030  OP use  0.93 (0.61)  0.95 (0.73)  –0.21  .831  ED use  0.16 (0.24)  0.21 (0.30)  –1.57  .119    Cancer (n = 143) n (%) or M (SD)  Noncancer (n = 119) n (%) or M (SD)  t test/chi-square statistics  p Value  Age (years)  54.7 (7.2)  45.9 (11.5)  7.23  <.001  Gender (F)  79 (55.2%)  69 (58.0%)  0.10  .749  CoC  85 (59.4%)  49 (41.2%)  7.96  .005  MCARE  57 (39.9%)  34 (28.8%)  3.17  .075  Comorbidity  2.42 (1.8)  2.39 (1.9)  0.11  .913  IP use  0.15 (0.43)  0.06 (0.09)  2.19  .030  OP use  0.93 (0.61)  0.95 (0.73)  –0.21  .831  ED use  0.16 (0.24)  0.21 (0.30)  –1.57  .119  IP inpatient; OP outpatient; ED emergency department; CoC Commission on Cancer accreditation affiliation; MCARE Medicare dual eligibility. View Large Table 2 Two group comparisons in cancer and noncancer groups   Cancer (n = 143) n (%) or M (SD)  Noncancer (n = 119) n (%) or M (SD)  t test/chi-square statistics  p Value  Age (years)  54.7 (7.2)  45.9 (11.5)  7.23  <.001  Gender (F)  79 (55.2%)  69 (58.0%)  0.10  .749  CoC  85 (59.4%)  49 (41.2%)  7.96  .005  MCARE  57 (39.9%)  34 (28.8%)  3.17  .075  Comorbidity  2.42 (1.8)  2.39 (1.9)  0.11  .913  IP use  0.15 (0.43)  0.06 (0.09)  2.19  .030  OP use  0.93 (0.61)  0.95 (0.73)  –0.21  .831  ED use  0.16 (0.24)  0.21 (0.30)  –1.57  .119    Cancer (n = 143) n (%) or M (SD)  Noncancer (n = 119) n (%) or M (SD)  t test/chi-square statistics  p Value  Age (years)  54.7 (7.2)  45.9 (11.5)  7.23  <.001  Gender (F)  79 (55.2%)  69 (58.0%)  0.10  .749  CoC  85 (59.4%)  49 (41.2%)  7.96  .005  MCARE  57 (39.9%)  34 (28.8%)  3.17  .075  Comorbidity  2.42 (1.8)  2.39 (1.9)  0.11  .913  IP use  0.15 (0.43)  0.06 (0.09)  2.19  .030  OP use  0.93 (0.61)  0.95 (0.73)  –0.21  .831  ED use  0.16 (0.24)  0.21 (0.30)  –1.57  .119  IP inpatient; OP outpatient; ED emergency department; CoC Commission on Cancer accreditation affiliation; MCARE Medicare dual eligibility. View Large Results of six regression analyses are summarized in Table 3. For patients with lung cancer, age and the number of comorbidities significantly predicted OP use and the number of comorbidities significantly predicted ED use. For non–lung cancer patients, only the number of comorbidities was the significant predictor of IP and ED uses. Explanatory variables include age, sex, whether the provider of the biopsy is a CoC-accredited organization, Medicare dual eligibility, and the number of comorbidities. Table 3 Regression model results   Cancer  Noncancer  Beta  SE  R2 change  p Value  Beta  SE  R2 change  p Value  IP use   Age  0.044  0.005  –.005  .400  –0.001  <0.001  –.016  .144   Sex (F)  –0.036  0.074  –.002  .628  –0.015  0.015  –.007  .311   CoC  0.039  0.074  –.002  .600  –0.026  0.016  –.020  .097   MCARE  –0.150  0.076  –.028  .051  –0.016  0.017  –.007  .336   Comorbidity  <0.001  0.021  <.001  .998  0.018  0.004  –.123  <.001  OP use   Age  –0.015  0.007  –.030  .028  0.010  0.006  –.022  .111   Sex (F)  0.032  0.097  –.001  .746  0.210  0.138  –.020  .130   CoC  –0.184  0.098  –.022  .063  0.044  0.143  –.001  .756   MCARE  –0.101  0.101  –.006  .316  –0.158  0.152  –.094  .303   Comorbidity  0.113  0.028  –.102  <.001  0.024  0.040  –.003  .547  ED use   Age  –0.003  0.003  –.008  .282  –0.005  0.003  –.027  .068   Sex (F)  –0.008  0.040  –.000  .841  0.038  0.055  –.004  .492   COC  –0.015  0.040  –.001  .710  –0.070  0.057  –.012  .220   MCARE  –0.042  0.042  –.007  .308  0.016  0.060  –.001  .787   Comorbidity  0.037  0.011  –.069  .002  –0.044  0.016  –.062  .006    Cancer  Noncancer  Beta  SE  R2 change  p Value  Beta  SE  R2 change  p Value  IP use   Age  0.044  0.005  –.005  .400  –0.001  <0.001  –.016  .144   Sex (F)  –0.036  0.074  –.002  .628  –0.015  0.015  –.007  .311   CoC  0.039  0.074  –.002  .600  –0.026  0.016  –.020  .097   MCARE  –0.150  0.076  –.028  .051  –0.016  0.017  –.007  .336   Comorbidity  <0.001  0.021  <.001  .998  0.018  0.004  –.123  <.001  OP use   Age  –0.015  0.007  –.030  .028  0.010  0.006  –.022  .111   Sex (F)  0.032  0.097  –.001  .746  0.210  0.138  –.020  .130   CoC  –0.184  0.098  –.022  .063  0.044  0.143  –.001  .756   MCARE  –0.101  0.101  –.006  .316  –0.158  0.152  –.094  .303   Comorbidity  0.113  0.028  –.102  <.001  0.024  0.040  –.003  .547  ED use   Age  –0.003  0.003  –.008  .282  –0.005  0.003  –.027  .068   Sex (F)  –0.008  0.040  –.000  .841  0.038  0.055  –.004  .492   COC  –0.015  0.040  –.001  .710  –0.070  0.057  –.012  .220   MCARE  –0.042  0.042  –.007  .308  0.016  0.060  –.001  .787   Comorbidity  0.037  0.011  –.069  .002  –0.044  0.016  –.062  .006  SE standard error; IP inpatient; OP outpatient; ED emergency department; CoC Commission on Cancer accreditation affiliation; MCARE Medicare dual eligibility. View Large Table 3 Regression model results   Cancer  Noncancer  Beta  SE  R2 change  p Value  Beta  SE  R2 change  p Value  IP use   Age  0.044  0.005  –.005  .400  –0.001  <0.001  –.016  .144   Sex (F)  –0.036  0.074  –.002  .628  –0.015  0.015  –.007  .311   CoC  0.039  0.074  –.002  .600  –0.026  0.016  –.020  .097   MCARE  –0.150  0.076  –.028  .051  –0.016  0.017  –.007  .336   Comorbidity  <0.001  0.021  <.001  .998  0.018  0.004  –.123  <.001  OP use   Age  –0.015  0.007  –.030  .028  0.010  0.006  –.022  .111   Sex (F)  0.032  0.097  –.001  .746  0.210  0.138  –.020  .130   CoC  –0.184  0.098  –.022  .063  0.044  0.143  –.001  .756   MCARE  –0.101  0.101  –.006  .316  –0.158  0.152  –.094  .303   Comorbidity  0.113  0.028  –.102  <.001  0.024  0.040  –.003  .547  ED use   Age  –0.003  0.003  –.008  .282  –0.005  0.003  –.027  .068   Sex (F)  –0.008  0.040  –.000  .841  0.038  0.055  –.004  .492   COC  –0.015  0.040  –.001  .710  –0.070  0.057  –.012  .220   MCARE  –0.042  0.042  –.007  .308  0.016  0.060  –.001  .787   Comorbidity  0.037  0.011  –.069  .002  –0.044  0.016  –.062  .006    Cancer  Noncancer  Beta  SE  R2 change  p Value  Beta  SE  R2 change  p Value  IP use   Age  0.044  0.005  –.005  .400  –0.001  <0.001  –.016  .144   Sex (F)  –0.036  0.074  –.002  .628  –0.015  0.015  –.007  .311   CoC  0.039  0.074  –.002  .600  –0.026  0.016  –.020  .097   MCARE  –0.150  0.076  –.028  .051  –0.016  0.017  –.007  .336   Comorbidity  <0.001  0.021  <.001  .998  0.018  0.004  –.123  <.001  OP use   Age  –0.015  0.007  –.030  .028  0.010  0.006  –.022  .111   Sex (F)  0.032  0.097  –.001  .746  0.210  0.138  –.020  .130   CoC  –0.184  0.098  –.022  .063  0.044  0.143  –.001  .756   MCARE  –0.101  0.101  –.006  .316  –0.158  0.152  –.094  .303   Comorbidity  0.113  0.028  –.102  <.001  0.024  0.040  –.003  .547  ED use   Age  –0.003  0.003  –.008  .282  –0.005  0.003  –.027  .068   Sex (F)  –0.008  0.040  –.000  .841  0.038  0.055  –.004  .492   COC  –0.015  0.040  –.001  .710  –0.070  0.057  –.012  .220   MCARE  –0.042  0.042  –.007  .308  0.016  0.060  –.001  .787   Comorbidity  0.037  0.011  –.069  .002  –0.044  0.016  –.062  .006  SE standard error; IP inpatient; OP outpatient; ED emergency department; CoC Commission on Cancer accreditation affiliation; MCARE Medicare dual eligibility. View Large Mortality risk Cox proportional hazard regression models (survival models) were developed for individuals with a lung cancer diagnosis (n = 143) and individuals without a lung cancer diagnosis (n = 119). For patients with lung cancer, significant factors include IP use per month and the number of comorbidities. Specifically, as shown in Table 4, individuals in the high IP use per month group (>0.06 encounters per month) are 2.5 times more likely to die (p < .001) than individuals in the low IP use per month group (≤0.06 encounters per month). However, individuals with a greater number of comorbid conditions (>2) were less likely to die (1–0.524 or 47.6%) (p = .005) than individuals with fewer comorbidities (1–2). Table 4 Cox proportional hazard model of time to death for individuals with a cancer diagnosis (n = 143) Factor  B  SE  Exp(B)  95% CI  p Value  Biopsy provider   Non-CoC provider (referent)      1       CoC provider  –0.23  0.22  0.79  [0.51–1.23]  .30  Age group   18–53 (referent)      1       54–64  0.29  0.23  1.34  [0.85–2.12]  .21  Sex   Male      1       Female  –0.22  0.22  0.80  [0.52–1.23]  .31  Inpatient (IP) utilization per month   Low (0.06 per month)      1       High (≥0.06 per month)  0.92  0.25  2.51  [1.55–4.07]  <.001  Number of comorbidities   Low (≤2)      1       High (>2)  –0.65  0.23  0.52  [0.33–0.83]  .005  Factor  B  SE  Exp(B)  95% CI  p Value  Biopsy provider   Non-CoC provider (referent)      1       CoC provider  –0.23  0.22  0.79  [0.51–1.23]  .30  Age group   18–53 (referent)      1       54–64  0.29  0.23  1.34  [0.85–2.12]  .21  Sex   Male      1       Female  –0.22  0.22  0.80  [0.52–1.23]  .31  Inpatient (IP) utilization per month   Low (0.06 per month)      1       High (≥0.06 per month)  0.92  0.25  2.51  [1.55–4.07]  <.001  Number of comorbidities   Low (≤2)      1       High (>2)  –0.65  0.23  0.52  [0.33–0.83]  .005  CoC Commission on Cancer, SE standard error, Exp(B) hazard ratio, CI confidence interval, degrees of freedom (df) = 1. View Large Table 4 Cox proportional hazard model of time to death for individuals with a cancer diagnosis (n = 143) Factor  B  SE  Exp(B)  95% CI  p Value  Biopsy provider   Non-CoC provider (referent)      1       CoC provider  –0.23  0.22  0.79  [0.51–1.23]  .30  Age group   18–53 (referent)      1       54–64  0.29  0.23  1.34  [0.85–2.12]  .21  Sex   Male      1       Female  –0.22  0.22  0.80  [0.52–1.23]  .31  Inpatient (IP) utilization per month   Low (0.06 per month)      1       High (≥0.06 per month)  0.92  0.25  2.51  [1.55–4.07]  <.001  Number of comorbidities   Low (≤2)      1       High (>2)  –0.65  0.23  0.52  [0.33–0.83]  .005  Factor  B  SE  Exp(B)  95% CI  p Value  Biopsy provider   Non-CoC provider (referent)      1       CoC provider  –0.23  0.22  0.79  [0.51–1.23]  .30  Age group   18–53 (referent)      1       54–64  0.29  0.23  1.34  [0.85–2.12]  .21  Sex   Male      1       Female  –0.22  0.22  0.80  [0.52–1.23]  .31  Inpatient (IP) utilization per month   Low (0.06 per month)      1       High (≥0.06 per month)  0.92  0.25  2.51  [1.55–4.07]  <.001  Number of comorbidities   Low (≤2)      1       High (>2)  –0.65  0.23  0.52  [0.33–0.83]  .005  CoC Commission on Cancer, SE standard error, Exp(B) hazard ratio, CI confidence interval, degrees of freedom (df) = 1. View Large Further, Figure 2 demonstrates the difference in hazard rate for individuals with a lung cancer diagnosis and biopsy performed by a CoC organization or affiliated provider compared with those having a biopsy performed by a non-CoC organization or affiliated provider. After 2 years, individuals with a biopsy performed by a CoC provider had approximately a 65% risk of death while those having had a biopsy by a non-CoC provider had approximately an 85% risk of death. Fig 2 View largeDownload slide Hazard function for individuals with a positive lung cancer biopsy. Fig 2 View largeDownload slide Hazard function for individuals with a positive lung cancer biopsy. The second Cox proportional hazards model included the same factors but considered only individuals without a lung cancer diagnosis. No factors were significant in this model. DISCUSSION Our findings must be placed in the context of this feasibility study to examine predictors of healthcare utilization and mortality for Medicaid individuals with or without a diagnosis of lung cancer. Our Medicaid dataset was limited to lung biopsy cases for 1 year and within a specific geographic (rural and urban) area of NY. Health utilization knowledge in low-income lung biopsy patients can provide an opportunity for providers and policy makers to improve care coordination for patients with lung cancer. Coordination of care can facilitate customized care interventions to address the specific needs of this population including surveillance following a diagnosis and SCPs for patients with lung cancer. The number and severity of comorbidities that coexist with a cancer diagnosis have been known to influence overall survival post-cancer diagnosis [22]. Our findings indicate that both the number of comorbidities and age are significant predictors of different types of healthcare services for patients with lung cancer, and only the number of comorbidities is a significant predictor of IP and ED use in non–lung cancer patients. Our findings report a higher utilization of ED and IP services among those with multiple comorbidities. Prognosis in lung cancer is determined by a myriad of factors in addition to the number of comorbidities [23]. Prior studies have proposed that mortality rates in metastatic cancer are a result of different factors including stage of disease (e.g., early vs. late), specific cancer biology, disparities in care, access to disease, or an early diagnosis in non-White patients [23, 24]. Our findings note that comorbidity influenced outcomes; however, we were not able to determine other factors because Medicaid claims data do not include stage of disease, race/ethnicity, and cancer-related biology information. In addition, we were not able to conclude whether socioeconomic status and insurance plans were contributing factors to health outcomes, as we did not compare different insurance plans. Our findings also noted that individuals with a greater number of comorbid conditions (two or more) were less likely to die than were individuals with fewer comorbidities (one to two). Our findings were not consistent with the literature on comorbidities and survival outcomes for patients with cancer [25]. The impact of all-cause survival and cancer-specific survival is likely to vary based on the type of cancer, stage of disease and prognosis, type of treatment and related sequela, and the number and severity of comorbidities [25]. Research is needed in the area of complex comorbidities in cancer and coordination of care to give clinicians opportunities for care coordination to reduce disease burden, health service utilization, and mortality rates. Our study outcome indicates that some factors related to healthcare utilization may be associated with receiving care at a CoC organization; however, with limited data we could not determine all factors related to healthcare utilization. Therefore, the specific role of the CoC accreditation status or non-CoC could not be determined as a predictor for outcomes. Fragmented survivorship care delivery is an important issue requiring new approaches to coordinated evidence-based care among those diagnosed with lung cancer [26]. As we report in our findings, patients with lung cancer used varied IP, OP, and ED healthcare services throughout the course of their illnesses. Stage of disease and symptom burden could influence the type and frequency of health service. For this study, we did not specifically track the type of each service encounter. The impact of healthcare service use on QOL was not measured in this study. Other studies have found, however, that care coordination to manage symptom burden and comorbidities is essential to improve QOL for patients with lung cancer [11]. Although our study did not stratify by stage of disease due to lack of availability in the Medicaid data or examine provider type of health service visits, it was noted that those patients with an association with a CoC organization at the time of biopsy had a greater chance of living longer than those patients associated with a non-CoC organization. Health-related issues influencing mortality in the noncancer population could be different from the health-related issues influencing mortality in the population of individuals who have a cancer diagnosis. Further, we have not yet identified in the data the reasons why Medicaid patients received a biopsy by a CoC versus a non-CoC organization. Our findings suggest that more research is needed to evaluate current quality standards such as the CoC to identify gaps and inform strategies to improve care coordination and healthcare quality. Important issues for future study will be comparisons across insurance types including employee-sponsored and publicly funded care plans as well as those individuals that would be less able to access surveillance following a diagnosis of cancer and cancer survivorship care [13, 14]. Many studies have reported disparities among patients with lung cancer including contributing factors such as race, unequal access to care, receipt of treatment, patient perception and trust, insurance status, and diagnosis at a later stage of disease [5, 8, 23, 27, 28]. Race/ethnicity, gender, socioeconomic status, and geography are significant contributors to lung cancer incidence [29]. Cancer incidence and mortality were reported higher among Black males and people of lower socioeconomic status [29, 30]. However, our Medicaid claims database did not have data on race/ethnicity. Linking cancer registry data to Medicaid claims data would address this gap in information. In addition, the average age for a lung cancer diagnosis is 65 years and older [31]. The median age for this study was 53 years. Few studies evaluate lung cancer in Medicaid patients, which insures younger and low income, relative to those in Medicare [32, 33]. Poor outcomes related to patients with lung cancer on Medicaid warrant future research to further evaluate predictors of health utilization and survival in this disparate population. Overcoming health disparities is key to improving care coordination for Medicaid patients with lung cancer. Research is needed to support ways to improve cancer detection, treatment, surveillance, QOL, and end of life care [11, 32, 33]. Limitations Several limitations for this study are noted. Although analysis of claims data has the benefit of including utilization across all settings for the vulnerable Medicaid population over multiple years, there are limitations. In the current analysis, the initial biopsy may not be the first lung biopsy and individuals may have had pre-existing lung cancer. When using a Medicaid database, it is not possible to determine the stage of cancer directly. Evidence-based guidelines (e.g., National Comprehensive Cancer Network) address all aspects of cancer disease management including screening, diagnosis, evaluation, staging, treatment, surveillance, and therapy [34]. This major limitation within the Medicaid data does not allow reliable comparisons between the treatments of patients with lung cancer and evidence-based guidelines. An overarching limitation of using a Medicaid database is the reliability for identifying cancer cases [35, 36]. Variability in cancer case identification is caused by several challenges including incomplete claims history caused by lack of continuous enrollment, age, missing procedural codes, and dual eligibility for Medicare and Medicaid programs [35, 36].In addition, challenges exist for the predictive ability of comorbidity and diagnostic-based indices, as well as how sensitive and specific these are in predicting outcomes related to comorbidity, mortality, and health utilization for patients with cancer [37]. Predictive ability of comorbidity indices can vary widely depending on the specific index (e.g., Charlson Comorbidity Index and Elixhauser Index), and selection of the appropriate index for use should consider type of available data, study population, and specific outcome(s) for the study [37]. Combining Medicaid data sources and the cancer registry data would provide a more complete and robust data, accounting for variables found exclusively in one or the other data set (e.g., stage of disease, race/ethnicity, socioeconomic status) [38]. Medicaid data provide coverage across populations when access to other databases such as SEER or a cancer registry database is not available [39, 40]. Therefore, linking to cancer registry data may be better option for combining data sources. Future studies linking cancer registry databases with Medicaid data would provide additional patient information including diagnosis, stage of disease, date of diagnosis, and cause of mortality providing opportunities to assess cancer care [41]. The average age for a lung cancer diagnosis is 65 years or older [31]; therefore, age may be considered a limitation for this study due to relatively fewer number of lung cancer cases in patients ≥18 to ≤64 years of age. Medicaid claims data can provide an opportunity to evaluate health service utilization and outcomes for future research directed toward a low-income and younger lung cancer population that may benefit from new treatments, health behavior change (e.g., tobacco cessation programs), and supportive care to improve QOL. The time for the calculation for mortality is short, and future research would benefit from a larger sample size and extended timeline for surveillance of data. The analysis is limited to the eight counties of Western New York where we have knowledge of affiliation with cancer centers. CONCLUSION Care coordination is essential for low-income individuals at risk or with a diagnosis of lung cancer. Patients in this population burdened with symptoms of their disease that affect overall QOL could benefit from appropriate coordination of care. Evidence-based lung cancer guidelines have been created to provide the most up-to-date information intended to improve the management of patient care as well as patient outcomes [2–9]. A recent series of articles published on cancer survivorship in the USA discussed how cancer survivorship is defined, the ongoing needs and preferences over the continuum of survivorship care, and considerations for coordinated cancer care [42–44]. To our knowledge, few studies have examined health service utilization of patients with lung cancer to explore opportunities for chronic care coordination based on type and frequency of service. Future research is needed to explore how patterns of health service utilization can be useful in developing and coordinating care plans across the cancer survivor continuum. Primary Data: Findings reported in this manuscript have not been previously published, and the manuscript is not being simultaneously submitted elsewhere. There has been no previous reporting of data; the authors have full control of all primary data, and the authors of this manuscript agree to allow the journal to review their data if requested. The authors have no study funding sources to report. Compliance with Ethical Standards Conflict of Interest: Authors Somayaji, Chang, Casucci, Xue, and Hewner declare that they have no conflict of interest. Ethical Approval: This article does not contain any studies with human participants or animals performed by any of the authors. Based on the data extraction and de-identification protocol, the University at Buffalo Institutional Review Board determined that this project did not qualify as human subjects research. Informed Consent: For this type of study, formal consent is not required. References 1. Williams DR, Kontos EZ, Viswanath Ket al.   Integrating multiple social statuses in health disparities research: The case of lung cancer. Health Serv Res . 2012; 47( 3 pt 2): 1255– 1277. Google Scholar CrossRef Search ADS PubMed  2. Alberg AJ, Brock MV, Ford JG, Samet JM, Spivack SD. Epidemiology of lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest . 2013; 143( 5 suppl): e1S– e29S. Google Scholar CrossRef Search ADS PubMed  3. McDonald KM, Sundaram V, Bravata DM,et al.   Closing the quality gap: A critical analysis of quality improvement strategies (vol. 7: care coordination). Rockville, MD: Agency for Healthcare Research and Quality (US); 2007 Jun. (Technical Reviews, No. 9.7.) Available from:https://www.ncbi.nlm.nih.gov/books/NBK44015/ 4. Rudin, Robert S. Courtney A. Gidengil, Zachary Predmore, Eric C. Schneider, James Sorace, and Rachel Hornstein, Identifying and Coordinating Care for Complex Patients: Findings from the Leading Edge of Analytics and Health Information Technology. Santa Monica, CA: RAND Corporation, 2016. Available from:https://www.rand.org/pubs/research_reports/RR1234.html 5. Slatore CG, Au DH, Gould MK; American Thoracic Society Disparities in Healthcare Group. An official American Thoracic Society systematic review: Insurance status and disparities in lung cancer practices and outcomes. Am J Respir Crit Care Med . 2010; 182( 9): 1195– 1205. Google Scholar CrossRef Search ADS PubMed  6. Mayer DK, Gerstel A, Walton ALet al.   Implementing survivorship care plans for colon cancer survivors. Oncol Nurs Forum . 2014; 41( 3): 266– 273. Google Scholar CrossRef Search ADS PubMed  7. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin . 2016; 66( 1): 7– 30. Google Scholar CrossRef Search ADS PubMed  8. Abdelsattar ZM, Hendren S, Wong SL. The impact of health insurance on cancer care in disadvantaged communities. Cancer . 2017; 123( 7): 1219– 1227. Google Scholar CrossRef Search ADS PubMed  9. Detterbeck FC, Lewis SZ, Diekemper R, Addrizzo-Harris D, Alberts WM. Executive summary: Diagnosis and management of lung cancer: American College of Chest Physicians evidence-based clinical practice guidelines. Chest . 2013; 143( 5 suppl): 7S– 37S. Google Scholar CrossRef Search ADS PubMed  10. Kaniski F, Enewold L, Thomas A, Malik S, Stevens JL, Harlan LC. Temporal patterns of care and outcomes of non-small cell lung cancer patients in the United States diagnosed in 1996, 2005, and 2010. Lung cancer . 2017; 103( suppl C): 66– 74. Google Scholar CrossRef Search ADS PubMed  11. Yang P, Cheville AL, Wampfler JAet al.   Quality of life and symptom burden among long-term lung cancer survivors. J Thorac Oncol . 2012; 7( 1): 64– 70. Google Scholar CrossRef Search ADS PubMed  12. Iyer S, Roughley A, Rider A, Taylor-Stokes G. The symptom burden of non-small cell lung cancer in the USA: A real-world cross-sectional study. Support Care Cancer . 2014; 22( 1): 181– 187. Google Scholar CrossRef Search ADS PubMed  13. Kehl KL, Liao K-P, Krause TM, Giordano SH. Access to accredited cancer hospitals within federal exchange plans under the Affordable Care Act. J Clin Oncol . 2017; 35( 6): 645– 651. Google Scholar CrossRef Search ADS PubMed  14. Graves JA, Swartz K. Effects of Affordable Care Act marketplaces and Medicaid eligibility expansion on access to cancer care. Cancer J . 2017; 23( 3): 168– 174. Google Scholar CrossRef Search ADS PubMed  15. Oncology ASoC. The state of cancer care in America, 2015: A report by the American Society of Clinical Oncology. J Oncol Pract . 2015; 11( 2): 79– 113. CrossRef Search ADS PubMed  16. Clarke JL, Bourn S, Skoufalos A, Beck EH, Castillo DJ. An innovative approach to health care delivery for patients with chronic conditions. Popul Health Manag . 2017; 20( 1): 23– 30. Google Scholar CrossRef Search ADS PubMed  17. McCabe MS, Bhatia S, Oeffinger KCet al.   American society of clinical oncology statement: Achieving high-quality cancer survivorship care. J Clin Oncol . 2013; 31( 5): 631– 640. Google Scholar CrossRef Search ADS PubMed  18. David EA, Cooke DT, Chen Y, Perry A, Canter RJ, Cress R. Surgery in high-volume hospitals not Commission on Cancer accreditation leads to increased cancer-specific survival for early-stage lung cancer. Am J Surg . 2015; 210( 4): 643– 647. Google Scholar CrossRef Search ADS PubMed  19. Cancer ACoSCo. Cancer Program Standards: Ensuring Patient-Centered Care , American College of Surgeons Commission on Cancer, Cancer Program Standards: Ensuring Patient-Centered Care, 2016 ed. 20. Bilimoria KY, Bentrem DJ, Stewart AK, Winchester DP, Ko CY. Comparison of commission on cancer-approved and -nonapproved hospitals in the United States: Implications for studies that use the National Cancer Data Base. J Clin Oncol . 2009; 27( 25): 4177– 4181. Google Scholar CrossRef Search ADS PubMed  21. H-CUP. Clinical Classifications Software for ICD-9-CM. 2016; Available at https://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp. Accessibility verified January 31, 2017. 22. Edwards BK, Noone AM, Mariotto ABet al.   Annual report to the nation on the status of cancer, 1975–2010, featuring prevalence of comorbidity and impact on survival among persons with lung, colorectal, breast, or prostate cancer. Cancer . 2014; 120( 9): 1290– 1314. Google Scholar CrossRef Search ADS PubMed  23. Grose D, Morrison DS, Devereux Get al.  ; Scottish Lung Cancer Forum. The impact of comorbidity upon determinants of outcome in patients with lung cancer. Lung Cancer . 2015; 87( 2): 186– 192. Google Scholar CrossRef Search ADS PubMed  24. Loh KP, Kansagra A, Shieh MSet al.   Predictors of in-hospital mortality in patients with metastatic cancer receiving specific critical care therapies. J Natl Compr Canc Netw . 2016; 14( 8): 979– 987. Google Scholar CrossRef Search ADS PubMed  25. Sarfati D, Koczwara B, Jackson C. The impact of comorbidity on cancer and its treatment. CA Cancer J Clin . 2016; 66( 4): 337– 350. Google Scholar CrossRef Search ADS PubMed  26. Klabunde CN, Haggstrom D, Kahn KL, Gray SW, Kim B, Liu B, Eisenstein J, Keating NL. Oncologists’ perspectives on post‐cancer treatment communication and care coordination with primary care physicians. Eur J of Cancer Care. 2017;26(4). 27. Yang R, Cheung MC, Byrne MMet al.   Do racial or socioeconomic disparities exist in lung cancer treatment? Cancer . 2010; 116( 10): 2437– 2447. Google Scholar PubMed  28. Niu X, Roche LM, Pawlish KS, Henry KA. Cancer survival disparities by health insurance status. Cancer Med . 2013; 2( 3): 403– 411. Google Scholar CrossRef Search ADS PubMed  29. Tabatabai MA, Kengwoung-Keumo JJ, Oates GRet al.   Racial and gender disparities in incidence of lung and bronchus cancer in the united states: A longitudinal analysis. Plos One . 2016; 11( 9): e0162949. Google Scholar CrossRef Search ADS PubMed  30. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2017. CA Cancer J Clin . 2017; 67( 1): 7– 30. Google Scholar CrossRef Search ADS PubMed  31. Society AC. Key Statistics for Lung Cancer. 2017; Available at https://www.cancer.org/cancer/non-small-cell-lung-cancer/about/key-statistics.html. Accessibility verified December 20, 2017. 32. Mack JW, Chen K, Boscoe FPet al.   Underuse of hospice care by Medicaid-insured patients with stage IV lung cancer in New York and California. J Clin Oncol . 2013; 31( 20): 2569– 2579. Google Scholar CrossRef Search ADS PubMed  33. Walker GV, Grant SR, Guadagnolo BAet al.   Disparities in stage at diagnosis, treatment, and survival in nonelderly adult patients with cancer according to insurance status. J Clin Oncol . 2014; 32( 28): 3118– 3125. Google Scholar CrossRef Search ADS PubMed  34. National Comprehensive Cancer Network. National Comprehensive Network Guidelines. 2018; Available at https://www.nccn.org/. Accessibility verified February 7, 2018. 35. Koroukian SM, Cooper GS, Rimm AA. Ability of medicaid claims data to identify incident cases of breast cancer in the Ohio Medicaid population. Health Serv Res . 2003; 38( 3): 947– 960. Google Scholar CrossRef Search ADS PubMed  36. Schrag D, Virnig BA, Warren JL. Linking tumor registry and medicaid claims to evaluate cancer care delivery. Health Care Financ Rev . 2009; 30( 4): 61– 73. Google Scholar PubMed  37. Yurkovich M, Avina-Zubieta JA, Thomas J, Gorenchtein M, Lacaille D. A systematic review identifies valid comorbidity indices derived from administrative health data. J Clin Epidemiol . 2015; 68( 1): 3– 14. Google Scholar CrossRef Search ADS PubMed  38. Schrag D, Virnig BA, Warren JL. Linking tumor registry and medicaid claims to evaluate cancer care delivery. Health Care Financ Rev . 2009; 30( 4): 61– 73. Google Scholar PubMed  39. Institute NC. List of SEER Registries. 2017; Available at https://seer.cancer.gov/registries/list.html. Accessibility verified December 22, 2017. 40. Sciences NCIDoCCP. National and State Cancer Registries. 2017; Available at https://epi.grants.cancer.gov/registries.html. Accessibility verified December 22, 2017. 41. Boscoe FP, Schrag D, Chen K, Roohan PJ, Schymura MJ. Building capacity to assess cancer care in the medicaid population in New York State. Health Serv Res . 2011; 46( 3): 805– 820. Google Scholar CrossRef Search ADS PubMed  42. Jacobs LA, Shulman LN. Follow-up care of cancer survivors: Challenges and solutions. Lancet Oncol . 2017; 18( 1): e19– e29. Google Scholar CrossRef Search ADS PubMed  43. Mayer DK, Nasso SF, Earp JA. Defining cancer survivors, their needs, and perspectives on survivorship health care in the USA. Lancet Oncol . 2017; 18( 1): e11– e18. Google Scholar CrossRef Search ADS PubMed  44. Nekhlyudov L, O’malley DM, Hudson SV. Integrating primary care providers in the care of cancer survivors: Gaps in evidence and future opportunities. Lancet Oncol . 2017; 18( 1): e30– e38. Google Scholar CrossRef Search ADS PubMed  © Society of Behavioral Medicine 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Translational Behavioral MedicineOxford University Press

Published: May 23, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off