Analysis of combined data from heterogeneous study designs: an applied example from the patient navigation research program.

University of South Florida Department of Family Medicine, Tampa, FL 33612, USA.
Clinical Trials (Impact Factor: 1.93). 01/2012; 9(2):176-87. DOI: 10.1177/1740774511433284
Source: PubMed


The Patient Navigation Research Program (PNRP) is a cooperative effort of nine research projects, with similar clinical criteria but with different study designs. To evaluate projects such as PNRP, it is desirable to perform a pooled analysis to increase power relative to the individual projects. There is no agreed-upon prospective methodology, however, for analyzing combined data arising from different study designs. Expert opinions were thus solicited from the members of the PNRP Design and Analysis Committee.
To review possible methodologies for analyzing combined data arising from heterogeneous study designs.
The Design and Analysis Committee critically reviewed the pros and cons of five potential methods for analyzing combined PNRP project data. The conclusions were based on simple consensus. The five approaches reviewed included the following: (1) analyzing and reporting each project separately, (2) combining data from all projects and performing an individual-level analysis, (3) pooling data from projects having similar study designs, (4) analyzing pooled data using a prospective meta-analytic technique, and (5) analyzing pooled data utilizing a novel simulated group-randomized design.
Methodologies varied in their ability to incorporate data from all PNRP projects, to appropriately account for differing study designs, and to accommodate differing project sample sizes.
The conclusions reached were based on expert opinion and not derived from actual analyses performed.
The ability to analyze pooled data arising from differing study designs may provide pertinent information to inform programmatic, budgetary, and policy perspectives. Multisite community-based research may not lend itself well to the more stringent explanatory and pragmatic standards of a randomized controlled trial design. Given our growing interest in community-based population research, the challenges inherent in the analysis of heterogeneous study design are likely to become more salient. Discussion of the analytic issues faced by the PNRP and the methodological approaches we considered may be of value to other prospective community-based research programs.

16 Reads
  • Source
    • "National outcome measures will be evaluated by the Patient Navigation Research Program [6,35,36]. Table 2 presents the patient characteristics of the study sample for our local analysis (n = 490). The majority was greater than 65 years of age (53%), African American (68%), and referred to our program based on an abnormal PSA test (84%); over one-third were diagnosed with prostate cancer. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Patient navigation programs have been launched nationwide in an attempt to reduce racial/ethnic and socio-demographic disparities in cancer care, but few have evaluated outcomes in the prostate cancer setting. The National Cancer Institute-funded Chicago Patient Navigation Research Program (C-PNRP) aims to implement and evaluate the efficacy of a patient navigation intervention for predominantly low-income minority patients with an abnormal prostate cancer screening test at a Veterans Affairs (VA) hospital in Chicago. Methods/Design From 2006 through 2010, C-PNRP implemented a quasi-experimental intervention whereby trained social worker and lay health navigators worked with veterans with an abnormal prostate screen to proactively identify and resolve personal and systems barriers to care. Men were enrolled at a VA urology clinic and were selected to receive navigated versus usual care based on clinic day. Patient navigators performed activities to facilitate timely follow-up such as appointment reminders, transportation coordination, cancer education, scheduling assistance, and social support as needed. Primary outcome measures included time (days) from abnormal screening to diagnosis and time from diagnosis to treatment initiation. Secondary outcomes included psychosocial and demographic predictors of non-compliance and patient satisfaction. Dates of screening, follow-up visits, and treatment were obtained through chart audit, and questionnaires were administered at baseline, after diagnosis, and after treatment initiation. At the VA, 546 patients were enrolled in the study (245 in the navigated arm, 245 in the records-based control arm, and 56 in a subsample of surveyed control subjects). Discussion Given increasing concerns about balancing better health outcomes with lower costs, careful examination of interventions aimed at reducing healthcare disparities attain critical importance. While analysis of the C-PNRP data is underway, the design of this patient navigation intervention will inform other patient navigation programs addressing strategies to improve prostate cancer outcomes among vulnerable populations.
    BMC Health Services Research 09/2012; 12(1):340. DOI:10.1186/1472-6963-12-340 · 1.71 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We evaluated the efficacy of a Chicago-based cancer patient navigation program developed to increase the proportion of patients reaching diagnostic resolution and reduce the time from abnormal screening test to definitive diagnostic resolution. Women with an abnormal breast (n = 352) or cervical (n = 545) cancer screening test were recruited for the quasi-experimental study. Navigation subjects originated from five federally qualified health center sites and one safety net hospital. Records-based concurrent control subjects were selected from 20 sites. Control sites had similar characteristics to the navigated sites in terms of patient volume, racial/ethnic composition, and payor mix. Mixed-effects logistic regression and Cox proportional hazard regression analyses were conducted to compare navigation and control patients reaching diagnostic resolution by 60 days and time to resolution, adjusting for demographic covariates and site. Compared with controls, the breast navigation group had shorter time to diagnostic resolution (aHR = 1.65, CI = 1.20-2.28) and the cervical navigation group had shorter time to diagnostic resolution for those who resolved after 30 days (aHR = 2.31, CI = 1.75-3.06), with no difference before 30 days (aHR = 1.42, CI = 0.83-2.43). Variables significantly associated with longer time to resolution for breast cancer screening abnormalities were being older, never partnered, abnormal mammogram and BI-RADS 3, and being younger and Black for cervical abnormalities. Patient navigation reduces time from abnormal cancer finding to definitive diagnosis in underserved women. Impact: Results support efforts to use patient navigation as a strategy to reduce cancer disparities among socioeconomically disadvantaged women. Cancer Epidemiol Biomarkers Prev; 21(10); 1691-700. ©2012 AACR.
    Cancer Epidemiology Biomarkers & Prevention 10/2012; 21(10):1691-700. DOI:10.1158/1055-9965.EPI-12-0535 · 4.13 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Patient navigation is a promising intervention to address cancer disparities but requires a multisite controlled trial to assess its effectiveness. Methods The Patient Navigation Research Program compared patient navigation with usual care on time to diagnosis or treatment for participants with breast, cervical, colorectal, or prostate screening abnormalities and/or cancers between 2007 and 2010. Patient navigators developed individualized strategies to address barriers to care, with the focus on preventing delays in care. To assess timeliness of diagnostic resolution, we conducted a meta-analysis of center-and cancer-specific adjusted hazard ratios (aHRs) comparing patient navigation vs usual care. To assess initiation of cancer therapy, we calculated a single aHR, pooling data across all centers and cancer types. We conducted a metaregression to evaluate variability across centers. All statistical tests were two-sided. Results The 10521 participants with abnormal screening tests and 2105 with a cancer or precancer diagnosis were predominantly from racial/ethnic minority groups (73%) and publically insured (40%) or uninsured (31%). There was no benefit during the first 90 days of care, but a benefit of navigation was seen from 91 to 365 days for both diagnostic resolution (aHR = 1.51; 95% confidence interval [CI] = 1.23 to 1.84; P <. 001)) and treatment initiation (aHR = 1.43; 95% CI = 1.10 to 1.86; P <. 007). Metaregression revealed that navigation had its greatest benefits within centers with the greatest delays in follow-up under usual care. Conclusions Patient navigation demonstrated a moderate benefit in improving timely cancer care. These results support adoption of patient navigation in settings that serve populations at risk of being lost to follow-up. © 2014 The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: [email protected] /* */
    JNCI Journal of the National Cancer Institute 06/2014; 106(6). DOI:10.1093/jnci/dju115 · 12.58 Impact Factor
Show more