In this Issue: Mixed Methods and Diverse Perspectives

CLINICAL AND POLICY INSIGHTS FROM MULTIMETHOD RESEARCH

This issue features 6 research articles that use a multimethod design^1–6 and an editorial⁷ that presents a typology of how to publish multimethod research. These studies show how qualitative and quantitative methods can inform and complement each other.

For example, the in-depth interview study by Rabago and colleagues¹ builds on the findings of a previous randomized clinical trial⁸ and pragmatic follow-up study.⁹ These quantitative studies showed the effectiveness of nasal irrigation for reducing sinus symptoms, decreasing use of sinus medications, and improving quality of life among patients with chronic sinusitis. In the current study, patients from the previous studies were asked to tell their stories. Qualitative analyses showed that the treatment empowered patients to control important aspects of their health and health care and improved their quality of life in diverse ways. Further analyses of the interview data uncovered barriers to use of nasal irrigation and strategies to overcome those barriers. By listening to the patient participant’s voice, this analysis offers clinicians and patients important information on how to implement an effective intervention and portrays the patient’s actual experience in ways that the quantitative aspects of the study cannot. By making the quantitative findings “real,” it empowers patients to understand what this treatment might mean for their lives, and it empowers clinicians to personalize therapy to individual patients.

The study by Wittink and colleagues² also follows up on a long line of ongoing investigation, in this case, of how older primary care patients experience depression. The study in this issue examines patients’ views of how interactions with their physicians influence the ways they communicate about depression. The findings provide important clues to how the interaction may affect the diagnosis of depression, and they may help to explain both underdiagnosis and differences in diagnosis by race, culture, and sex.

Companion articles by Solberg and colleagues³ and Hroscikoski and colleagues⁴ report the quantitative and qualitative findings, respectively, of a multimethod study. This study evaluated implementation of the Chronic Care Model, which is being widely used to guide practice and health care system reform.^10–12 The quantitative findings,³ limited by small sample size and a pre-post design, show improvements in care for 3 chronic illnesses; however, only 2 Chronic Care Model elements—clinical information systems and decision support—are associated with improvements in care for one of the chronic illnesses, diabetes. The qualitative component of the study⁴ reveals the meaning of the quantitative findings. By using multiple data sources to examine 5 of 18 participating practices in depth, this analysis shows the health care system and practice contexts for the intervention—essential information for others attempting to translate the findings into their practice settings. These findings show that the limited initial effect of the intervention caused participants to change their strategies in ways that did not build on the strengths of previously successful change interventions. Findings about competing priorities and the challenges of specifying components of the model should be useful to others attempting to implement or adapt the Chronic Care Model to improve care.

The study by Sussman and colleagues⁵ mixes interviews, focus groups, and a survey to take an in-depth look at the decision to undertake preventive counseling for obesity. The findings from this on-the-ground perspective in real-world practices provide a welcome complement to previous health services research. Earlier research has provided the view from 10,000 feet and, as a result, has yielded simplistic solutions to a problem that this study finds to be complex. The identification of stable factors that set the stage for counseling, coupled with dynamic factors during the visit, provide a framework for more creative and realistic interventions in the future.

Bennett and colleagues⁶ use a quantitative sampling frame to identify women with inadequate prenatal care, and then they follow up with individual interviews and focus groups. They found that communication with clinicians is an important factor influencing prenatal care use. This potentially mutable factor and the specific ideas about “breaking it down” point to novel next steps for interventions to improve prenatal care use.

OTHER ARTICLES OF CLINICAL AND POLICY INTEREST

Many studies show that patients’ pain is poorly managed at all levels of the health care system. In a randomized clinical trial in 14 rural primary care practices, Ahles and colleagues show that patients with pain and psychosocial problems improve with telephone assistance by a nurse-educator.¹³ The intervention appears to be well integrated into primary care, improving not only patients’ bodily pain but also role function, vitality, and emotional function compared with usual care. The sustained effect at 12 months is remarkable, considering that the average patient received only 3 telephone calls.

Recent concerns about bioterrorism have generated interest in strategies for early identification of symptom complexes that could be associated with various illnesses. A study in a primary care practice finds that syndromic surveillance using billing data is feasible and shows earlier identification of an increase in cases of influenza than do data from emergency departments in 5 adjacent counties.¹⁴ It appears that billing data from primary care practice can be used to identify emerging infections in the community. With a bit of ongoing infrastructure support, practice networks can conduct ongoing surveillance to meet a number of needs. This idea is not new,¹⁵ but this proof of its feasibility and possible superiority to emergency department surveillance argues for the development of sustained support systems.

In a provocative study of a nationally representative sample, Viera and colleagues find that 20% of US adults do not have a usual source of care.¹⁶ For most people, this lack of a medical home is due to their preference. It is also, however, associated with lack of insurance, male sex, excellent reported health, younger age, and Hispanic ethnicity. This study can help to identify those who are excluded from a medical home vs those who choose to be medically homeless and can lead to targeted interventions to foster access to a medical home.

The essay in this issue¹⁷ makes the case that house calls are not a relic of family medicine’s past but a potential key to the future. Landers finds house call practice to be financially viable, as well as satisfying for both patients and clinicians.

We encourage readers to join the Annals online discussion at http://www.AnnFamMed.org.