A prescription for publication

Editor’s note: Jan S. Redfern is president of Redfern Strategic Medical Communications Inc., Goshen, N.Y. Laura F. Quinn is senior vice president, specialisms and consultative engagements, at Synovate Healthcare, a Mahwah, N.J., research firm.

For the most part, qualitative or quantitative market research is, understandably, confidential and proprietary. However, it is becoming increasingly common for pharmaceutical and health care companies to express an interest in publishing a patient or physician attitudinal survey in a peer-reviewed journal. Provided that the survey is formulated with a high degree of scientific merit and presented with objectivity, this type of publication offers valuable information for health care providers and a credible citation for marketers.

In this context, this article will furnish practical guidance on the design, execution and analysis of patient, health care provider or general-public surveys appropriate for journal publication; and provide insight into the publication process.

The commentary is based on a consolidation of the authors’ collective experiences in designing and implementing market research, coauthoring research articles and reviews and successfully publishing surveys in peer-reviewed journals.

Commit early

One of the most important aspects of publishing survey findings relates to the scope and types of questions asked. To enhance the likelihood of journal acceptance, it is prudent to commit to a publication early in the questionnaire development process rather than deciding to publish after the survey is completed. In this way, the survey questions can be focused on issues of contemporary clinical or scientific importance and oriented toward awareness, attitudes, knowledge, behaviors and treatment practices. Questions of a more product-specific nature can be partitioned in a separate section of the survey that is pre-specified as proprietary or confidential. Industry standard procedures should be followed to ensure the absence of bias in the wording of questions and in the order or position of questions in the questionnaire.

Examples of topics that may interest peer-review journals are summarized in Table 1.

These examples are based, in part, on successfully published surveys in which Synovate Healthcare or Redfern Strategic participated either in the design and execution1-3 or in the facilitation of publication.4-6

It is critical that the author(s) of the article provide input on all aspects of the survey from the onset (Figure 1) and meet the requirements for authorship espoused by the International Committee of Medical Journal Editors.7 If necessary, the author(s) should also obtain approval from their institutional review board to conduct the survey.

Require a thorough accounting

Another pivotal aspect of survey methodology relevant to publication is the method used to select survey participants randomly once the population of interest (general public, patients, health care providers, etc.) has been defined. Journals typically require a thorough accounting of the flow of participants through each stage of the survey (ideally presented in the form of a consort diagram) and should include the total number of participants who were initially approached, the number of screened participants who were disqualified from taking the survey and the reasons for disqualification, and finally, the proportion of screened participants who completed the survey.

Specific details on the manner of random selection are essential to bolster confidence in the survey findings and ensure that the survey sample is either a true reflection of the population universe or that screening criteria are purposefully over-recruiting in order to emphasize a population of particular interest (e.g., a particular disease subtype).
An article based on a multinational, general-public survey, for example, should adequately describe a number of methodological issues: 

  • The basis for random selection of participants for telephone (e.g., random-digit dialing) or face-to-face interviews or from in-house panels.
  • Weighting methods used to balance results to known universe parameters (e.g., age, race, gender and education).
  • Clustering techniques whereby major cities or regions are segmented into several zones and participants are randomly chosen within a zone.
  • The rationale for country selection.
  • The impact (if any) on survey execution of local customs/sensibilities and the availability of and access to telecommunication services.
  • Any incentives offered to encourage participation in the survey (e.g., gifts or cash, enrolment in sweepstakes and prize drawings).

Journals will also typically require evidence that the survey has been appropriately tested in pilot studies with actual respondents to ensure that questions are easily understood and interpreted similarly among respondents (Figure 1). Appropriate validation is also necessary to make sure that the survey measures what it is intended to measure and shows a high level of internal reliability (consistent results across similar, related items).

In addition, it should be kept in mind that the validity of any global survey is strongly impacted by local translations of the specific terms of interest; this is especially important in some regions such as Latin America, where substantial diversity exists in both culture and language. A complete description of the translation process is required and should encompass the use of native professional translators to convert the questionnaire and back-check the translations to ensure quality control.

Solid statistical foundation

It is essential to apply sound statistical principles to determine the number of respondents to participate in the survey. This is an extremely important consideration in the early stages of planning, and journals will expect to see that the survey is grounded on a solid statistical foundation rather than influenced solely by budgetary or other issues. The sample size should be large enough to allow accurate conclusions (especially if analysis of subgroups is involved) but not too large that cost effectiveness is negatively impacted.
The calculation of survey sample size for a population-based survey should take into account:

  • the prevalence of the variable of interest;
  • the desired confidence level (the level of uncertainly tolerate - typically 95 percent) and margin of error (typically 5 percent); and
  • the expected response distribution.

A majority of surveys utilize nothing more than simple descriptive statistics (means, ranges, standard deviations, etc.) to aid interpretation. However, the absence of formal statistical analysis substantially weakens the survey findings and ultimately jeopardizes publication acceptance. Most journals expect some kind of statistical testing appropriate for the type of data being analyzed (categorical, continuous, ordinal, etc.). Typically, journals also require a statement regarding the margin of error. However, it is important to note that it is not statistically valid to provide a margin of error for a survey in its entirety. Margin of error applies only to a specific question and is impacted not only by sample size and confidence interval but also by questionnaire structure. Consultation with a statistician is highly recommended to select the most appropriate type of analysis and provide input on the margin of error calculation.

Finally, it is important to provide details of any weighting methods that are employed to generalize survey results to the respective populations as well as description of safeguards to ensure data integrity and accuracy (e.g., removal of fraudulent or duplicate respondents, etc.).

Can help streamline

Partnering with a publications scientist with journal experience can help streamline all stages of the manuscript development process. This includes, for example, selecting an appropriate journal, researching the topic, facilitating manuscript development and communicating with consultants and journal editors.

It is unusual for a survey of any type to be accepted by a top-tier journal (The Journal of the American Medical Association, The New England Journal of Medicine, The Lancet, etc.), unless, of course, the survey is co-sponsored by a major medical organization affiliated with the journal. Consequently, submission to a second-tier journal may be a more realistic approach. Depending on the focus of the survey and the population studied, the article can be submitted to a specialist journal (e.g., a gerontology- or pediatric-oriented journal for a survey involving the elderly or children) or to a journal catering to a more general medical audience.

Choosing a target journal is a critical consideration and is influenced by a number of issues:

  • the quality, relevance and clinical significance of the survey findings in relation to the journal readership and mission statement;
  • the desired time frame for publication; and
  • the journal’s circulation and impact factor.

If there is any uncertainty about the potential interest of a journal in the survey manuscript, it is appropriate to write to the editor (prior to submitting for consideration of publication) briefly describing the purpose of the survey and explaining why the study findings may be of interest to journal readers. Early feedback from the editor on the suitability of the article will save considerable time by excluding journals that are only vaguely or definitely not interested in publishing the article.

The article should discuss the survey findings in the context of existing medical or scientific literature and emphasize the clinical significance and implications of the results. In addition, it is appropriate to present any limitations of the survey that may potentially impact interpretation.

The acknowledgement section of the article must include a full disclosure of the contribution of the author(s) and other individuals who were involved in the conception, design, analysis, interpretation and writing of the survey. In addition, it should clearly delineate any financial support for the research and manuscript development.

The final step in the publication process is addressing editors’ and reviewers’ comments. Unless the article is completely rejected by the journal, the chance of publication is very good provided the comments can be adequately addressed. The skill in this process is knowing when to make the requested changes and when to push back without jeopardizing acceptance. It is not necessary to make every change requested but it is important to provide a sound rationale why specific changes are inappropriate, impractical or impossible to accommodate.

Valuable and citable resource

Peer-reviewed publication of patient, health care provider or general-public survey data creates a valuable and citable resource for health care providers. However, in the current climate of increasing regulatory and legal scrutiny, it is advisable that surveys destined for publication are prospectively designed and based on an amalgam of scientific merit, integrity and transparency.

The keys to successfully publishing a peer-reviewed journal article based on survey findings include the following:

  • Commit to a publication early in the questionnaire development process and obtain author(s) input on all aspects of the survey from the onset.
  • Focus survey questions on issues of contemporary clinical or scientific importance.
  • Apply sound statistical principles to determine sample size and analyze results.
  • Choose an experienced publications scientist to streamline manuscript development and submission.
  • Present the survey findings with objectivity and discuss in the context of existing medical or scientific literature, emphasizing the clinical significance of the results.
  • Fully disclose the contribution of all individuals involved in the survey and clearly delineate any financial support for the research and manuscript development.

References

1 Keye W.R., Jr., Bradshaw K.D. “A Survey of the Practices and Opinions of the Domestic Members of The American Society for Reproductive Medicine.” Fertility and Sterility. 2004 Sept.;82(3):536-42.

2 Hann H.W., Han S.H., Block T.M., Harris M., Maa J.F., Fisher R.T., et al. “Symptomatology and Health Attitudes of Chronic Hepatitis B Patients in the USA.” Journal of Viral Hepatitis. 2008 Jan.;15(1):42-51.

3 Sekeres M.A., Schoonen W.M., Kantarjian H., List A., Fryzek J., Paquette R., et al. “Characteristics of U.S. Patients With Myelodysplastic Syndromes: Results of Six Cross-Sectional Physician Surveys.” Journal of the National Cancer Institute. 2008 Nov. 5;100(21):1542-51.

4 Paek E., Johnson R. “Public Awareness and Knowledge of Herpes Zoster: Results of a Global Survey.” Gerontology. 2009;56(1):20-31.

5 Cash T.F. “Attitudes, Behaviors and Expectations of Men Seeking Medical Treatment For Male Pattern Hair Loss: Results Of A Multinational Survey.” “Attitudes and Practices of Dermatologists and Primary Care Physicians Who Treat Patients for MPHL: Results of a Survey.” Current Medical Research and Opinion. 2009 July. Feb.;25(7):1811-20.

6 Cash T.F. “Attitudes and Practices of Dermatologists and Primary Care Physicians Who Treat Patients for MPHL: Results of a Survey.” Current Medical Research and Opinion. 2010 Feb.;26(2):345-54.

7 International Committee of Medical Journal Editors. The Uniform Requirements for Manuscripts Submitted to Biomedical Journals. Publication Ethics: Sponsorship, Authorship and Accountability. www.icmje.org/urm_fullpdf, accessed June 21, 2010.