Elsevier

Journal of Surgical Research

Volume 224, April 2018, Pages 112-120
Journal of Surgical Research

Research Methods Paper
Nonresponse bias in survey research: lessons from a prospective study of breast reconstruction

https://doi.org/10.1016/j.jss.2017.11.058Get rights and content

Abstract

Background

Survey-based research is essential for evaluating the outcomes of health care in an era of patient-centered care. However, many such studies are hampered by poor response rates in completion of study questionnaires, thus limiting the generalizability of any findings. The objectives of this analysis were to identify independent variables associated with nonresponse to surveys following breast reconstruction to improve future patient-reported outcomes research.

Materials and methods

The Mastectomy Reconstruction Outcomes Consortium is a prospective cohort study involving 11 leading medical centers from the United States and Canada. Nonresponse rates for surveys assessing satisfaction with breast, satisfaction with care (BREAST-Q), depression (Patient Health Questionnaire-9), and anxiety (Generalized Anxiety Disorder-7) were measured at 1 y and 2 y postoperatively. Clinical complication rates were compared between responders and nonresponders, and multivariable models were used to assess predictors of nonresponse.

Results

Among 2856 women in the analytic cohort, 1882 (65.9%) underwent implant-based, 817 (28.6%) received autologous, and 157 (5.5%) underwent latissimus dorsi myocutaneous flap breast reconstructions. Nonresponse rates to surveys at 1 y and 2 y were 27.8% and 34.4%, respectively. Race, ethnicity, and annual household income were associated with nonresponse to surveys. Women who underwent implant-based procedures were less likely to complete long-term surveys.

Conclusions

As survey-based research plays an increasingly prominent role in evaluating the outcomes of breast reconstruction, we found socioeconomic and procedure-related differences in survey response rates. Investigators must consider systematic differences in response rates among particular groups of women on the generalizability and validity of findings and perform rigorous nonresponse bias analyses.

Introduction

Advances in the measurement of patient-reported outcomes (PROs) have enabled surgeons and health services researchers to investigate improvements in health-related quality of life related to surgical care. Furthermore, as comparative effectiveness research becomes increasingly prominent in nearly all areas of surgery, survey-based research methods are also frequently used to compare the impact of different surgical procedures. As we move forward in this era of patient-centered care, it has become increasingly clear that survey-based research is an essential component of assessing the impact of surgical procedures.

The American Cancer Society estimates that breast cancer is the most common nonskin cancer among women in the United States.1 Although breast conservation remains the primary surgical treatment for breast cancer, mastectomy is still performed on a routine basis.2, 3, 4, 5 To decrease the adverse effects of mastectomy on psychosocial functioning, many women undergo breast reconstruction after mastectomy. According to the American Society for Plastic Surgeons, approximately 109,256 women underwent breast reconstruction after mastectomy in the United States in 2016.6 Building on recent advances in survey-based research, surgeons and health services researchers commonly use validated PRO instruments to evaluate health-related quality of life and psychosocial functioning among women undergoing breast reconstruction after mastectomy.7, 8, 9, 10, 11, 12

In contrast to medical records or claims data, measuring PROs is entirely dependent on the patients' willingness to complete surveys. Importantly, systematic differences between patients who respond to surveys and those who do not respond may lead to nonresponse bias, which as a form of measurement bias threatens the generalizability and validity of study findings. Although PRO measures have been increasingly applied in breast reconstruction and other areas of surgery, few authors have assessed or adequately adjusted for nonresponse bias in study findings.

Investigators should not assume that nonresponse to surveys occurs randomly. Patients' willingness to complete postoperative surveys may be affected by patient, clinical, or procedure-related differences. To assess patient outcomes in a rigorous manner, investigators should perform nonresponse bias analyses. Using data collected in a multiinstitutional, prospective cohort study of breast reconstruction, our objective in the current analysis was to identify sociodemographic, clinical, and procedural characteristic associated with not completing postoperative surveys. In this article, we also discuss approaches to assessing and adjusting for nonresponse bias for PRO surveys in this study population.

Section snippets

Data source and analytic cohort

The Mastectomy Reconstruction Outcomes Consortium Study is a 5-year National Cancer Institute–funded, longitudinal, prospective cohort study assessing clinical and PROs of breast reconstruction after mastectomy. This study enrolled eligible patients at 11 participating centers in the United States and Canada between February 2012 and July 2015. Institutional approval was obtained at all participating institutions, and informed consent was obtained from all women who participated in the study.

Results

Among 2856 women enrolled in the Mastectomy Reconstruction Outcomes Consortium Study, 1882 (65.9%) underwent implant-based, 817 (28.6%) received autologous, and 157 (5.5%) underwent latissimus dorsi myocutaneous flap breast reconstruction. Nonresponse rates to BREAST-Q surveys at 1 and 2 y postoperatively were 27.8% and 34.4%, respectively (Table 1). Nonresponse rates to PHQ/GAD surveys at 1 and 2 y postoperatively were 29.3% and 36.4%, respectively (data not shown). At 1 y after

Discussion

Nonresponse bias is an ongoing concern for investigators in survey-based research because not all study participants will complete postoperative surveys. Using data collected from the Mastectomy Reconstruction Outcomes Consortium, we demonstrate that nonresponse to surveys did not occur randomly among women who underwent breast reconstruction. We found important socioeconomic effects on survey response rates that allude to an incomplete understanding of previously reported disparities in the

Conclusion

In this era of patient-centered care and precision medicine initiatives, rigorous qualitative and quantitative studies will build upon our current understanding of the impact of surgical procedures. As survey-based research continues to play an increasingly prominent role in breast reconstruction and other areas of surgery, our study demonstrates important socioeconomic disparities and procedure-related differences in survey response rates. Investigators assessing PROs after breast

Acknowledgment

Authors' contributions: Berlin, Hamill, Qi, Kim, and Wilkins conceptualized and designed the study. Berlin, Hamill, Qi, Kim, Pusic, and Wilkins performed the acquisition, analysis, or interpretation of data. Berlin drafted the manuscript. Qi and Kim performed the statistical analysis. All authors gave critical revision of the manuscript for important intellectual content.

Funding: This study was supported by grants from the National Cancer Institute (1RO1CA152192).

Conflict of interest: The

References (29)

  • E.B. Habermann et al.

    Are mastectomy rates really increasing in the United States?

    J Clin Oncol

    (2010)
  • Reconstructive Plastic Surgery Statistics 2015

  • E.G. Wilkins et al.

    Prospective analysis of psychosocial outcomes in breast reconstruction: one-year postoperative results from the Michigan breast reconstruction outcome study

    Plast Reconstr Surg

    (2000)
  • L. Stevens et al.

    The psychological impact of immediate breast reconstruction for women with early breast cancer

    Plast Reconstr Surg

    (1984)
  • Cited by (29)

    • Access to Results of Patient Reported Outcome Surveys Does Not Improve Survey Response Rates

      2023, Journal of Surgical Research
      Citation Excerpt :

      Other studies have shown an even more significant decrease in longitudinal PRO survey response rate beyond 2 y from baseline.19 Diminishing response rates significantly limit the ability to track longitudinal changes in PROs and may increase nonresponse bias which may limit the validity and generalizability of findings from long-term PRO studies.20 These limitations to longer-term studies may lead clinicians to place greater focus on treatments and interventions that improve short-term satisfaction at the expense of long-term QoL.21

    • Development of an evidence-based approach to the use of acellular dermal matrix in immediate expander-implant-based breast reconstruction

      2021, Journal of Plastic, Reconstructive and Aesthetic Surgery
      Citation Excerpt :

      Finally, as is commonly the case with survey studies, MROC experienced a sizable drop rate of approximately 40% at two years following reconstruction. Although we adjusted for non-response bias, it remains possible that survey nonresponders could have affected the study findings.37 In an attempt to provide an evidence-based approach to the use of ADM in immediate expander-implant-based breast reconstruction, this study aimed to identify patient subgroups within which ADM might reduce risks or improve PROs for these procedures.

    • Participation Bias in a Survey of Community Patients With Heart Failure

      2020, Mayo Clinic Proceedings
      Citation Excerpt :

      Reports on the association between participation and sex varied markedly depending on the survey topic, year, and target population.22-24 Consistent with studies showing that being unmarried is associated with nonparticipation, we also observed that nonparticipants were more likely to be single than participants23,25-27 and have lower education levels.23 These results may indicate that the underrepresentation of unmarried or less educated individuals may cause nonresponse bias if being single or having a lower education level is associated with the survey measures.

    View all citing articles on Scopus

    This study was presented at the annual meeting of the Plastic Surgery Research Council in Durham, North Carolina on May 6, 2017.

    View full text