Primary care physicians’ perceived barriers and facilitators to conservative care for older adults with chronic kidney disease: design of a mixed methods study

Mixed methods study design

Mixed methods research involves the following: identification of a specific mixed
method research design; data collection and analysis of qualitative and quantitative
data; integration of the two forms of data; and assignment of prioritization to either
or both forms of data 6]. A sequential exploratory mixed methods design (Fig. 1), a two-phase study design, will be conducted for the purposes of this study. The
first phase of the study will involve qualitative description using interviews to
understand and describe primary care physicians’ perspectives of care for older adults
(aged 75 years and greater) with stage 5 CKD (estimated glomerular filtration rate
(eGFR) less than 15 mL/min/1.73 m
²
for at least 3 months) not on dialysis in the primary care setting. A subsequent quantitative
phase (phase 2) will be designed to quantify and describe the magnitude of barriers
and facilitators to conservative care. The survey will be used to also examine the
relationship between demographic and clinical practice characteristics and barriers
and facilitators to care. The quantitative phase will involve questionnaire design,
testing, and administration. The qualitative and quantitative phases are assigned
equal priority and the two phases will be integrated sequentially (i.e., the quantitative
phase will build on the qualitative phase). A study in the field of nephrology using
a similar type of design has been previously reported 7]. Ethical approval of this mixed methods study has been obtained from the Conjoint
Health Research Ethics Board at the University of Calgary.

Fig. 1. The mixed methods study design

Phase 1: qualitative interviews

Qualitative paradigm

Primary care physicians’ perspectives of barriers and facilitators to caring for older
adults with stage 5 CKD are complex and incorporate individual views and experiences
of diagnosis, treatment, and referral. Hence, the nature of this research question
encompasses multiple truths and rich perspectives. These multiple perspectives are
best captured through the collection of data in the natural setting or context where
the participants experience the barriers and facilitators 8]. Ultimately, the qualitative paradigm is ideal for this component of the research,
as it allows for a comprehensive and nuanced account of the problem under study. Fundamental
qualitative description 9] will be employed to develop a comprehensive summary of barriers and facilitators
experienced by physicians when providing care to older adults with stage 5 CKD.

Participant selection and recruitment

Using a snowball sampling approach, key stakeholders or opinion leaders in areas of
family medicine, seniors’ health, and kidney disease will be asked to identify primary
care physicians with experience caring for older adults with stage 5 CKD in the community.
Eligible participants will include physicians who currently practice in the primary
care setting, have at least 1 year of experience as a primary care physician, and
have experience caring for patients 75 years or older with stage 5 CKD in the past
year. Consenting physicians will be interviewed in-person or via telephone, based
on their preference. Participants will also be asked to recommend other providers
who meet the eligibility criteria.

Data collection

Participants will be asked to provide basic information regarding demographics and
practice characteristics. An interview guide has already been developed based on a
review of the literature and discussion with the research team, including key decision-makers.
The broad introduction question, probing questions, and closing question will be pilot
tested with three primary care physicians (see Additional file 1). The interview guide considered barriers, facilitators, and strategies regarding
the following overarching topics: stage 5 CKD diagnosis and assessment, care management,
and resource use. One interviewer (HT) will conduct the interviews to ensure consistency
in interview style and structure. Participant recruitment will cease when theoretical
saturation is reached (i.e., when new information or concepts no longer emerge from
additional interviews). All interviews will be audio-recorded and then transcribed
verbatim by a professional transcriptionist.

During the semi-structured interviews, providers will be asked about their perspectives
on caring for the population of interest. The probing questions are designed to help
interviewees think about barriers they experience regarding diagnosis, care management,
and referral decisions, should these topics not develop spontaneously. Participants
may also be asked about potential strategies or facilitators they could employ to
address their challenges. All participants will be asked the broad introductory and
closing questions, with probing questions used only when necessary. Field notes will
also be documented on the participants’ responses to interview questions, which may
inform data analysis and subsequent conduct of interviews.

Qualitative analysis

Perceptions relating to diagnosis and management decision-making will be identified
and categorized using conventional content analysis 10], a method of interpreting interview data with the goal of describing the phenomenon
of interest. The steps for conventional content analysis include the following: (1)
achieving immersion by first reading the interview data in its entirety to acquire
an overall sense of the phenomenon; (2) reading interviews word by word and highlighting
words that capture key concepts that become codes; (3) documenting initial impressions,
thoughts, and interpretation; (4) developing codes; and (5) sorting codes that are
related to each other into themes and sub-themes. Definitions will be developed for
existing codes, themes, and sub-themes, and exemplars of these will be reported in
the findings. Coding will be conducted in triplicate (i.e., there will be three independent
researchers coding each of the manuscripts) and then reviewed with the research team.

The process will be reflexive and interactive as continual data collection and data
analysis will shape each other 9]. For example, code titles or definitions identified based on the first interview
may be modified based on data collected during the second interview, and new codes
may be added requiring recoding of the first interview transcript. Codes will be generated
from the interview data, and they will be systematically applied to identify themes
and patterns. Results from the qualitative phase of this mixed methods study has been
published elsewhere 11]. These results will facilitate item generation for survey development in the next
study phase, as the codes, themes, and sub-themes will directly form the variables
of interest in the questionnaire (phase 2).

Phase 2: quantitative survey

Participant selection

The sampling frame for this population-based survey will include all primary care
physicians from the College of Physicians and Surgeons of Alberta, a provincial regulatory
body of medical practice for physicians 12]. Employing a series of screening questions on the cover letter of the questionnaire,
physicians will be asked to complete the questionnaire only if they recall ever having
experience providing care to people 75 years or greater with an eGFR of less than
15 mL/min/1.73 m
²
. Providers will be invited to participate via mail, fax, and online methods.

Questionnaire development and administration

Based on results from the qualitative phase of the study, the questionnaire will obtain
more information about barriers and facilitators to provision of care for older adults
with stage 5 CKD in the community and will include a combination of multiple choice,
Likert scales, and open-ended responses. The questionnaire is likely to reflect diagnosis,
management, and referral decision-making components of care for older patients with
stage 5 CKD. Potentially sensitive items focused on personal attributes and clinical
practice settings will be situated at the end of the questionnaire and will include
the following: the respondent’s age; gender; practice interests (e.g., care of the
elderly) 13]; years practicing in this specialty; time allocated for clinical practice; and practice
in an urban or rural area based on population. Previous studies have demonstrated
that personal attributes and practice environments are key predictors of behavior
among primary care physicians 14]. The questionnaire will be developed in two steps: item identification and preliminary
development of the questionnaire and questionnaire refinement.

Item identification and preliminary development of the questionnaire

Potential items for inclusion in the questionnaire will be informed via a literature
review and interviews from phase 1. Hence, addressing the qualitative objective will
provide specific items relating to primary care for the patient population of interest,
which will be categorized into different domains according to the Theoretical Domains
Framework 15], 16]. The Theoretical Domains Framework provides a strategy for assessing problems related
to implementation of clinical practice guidelines and provides a foundation for intervention
development. It has been validated and demonstrated to be useful over a range of different
healthcare systems, containing the following 14 domains that may influence behavior
and provide a comprehensive structure for categorizing barriers and facilitators to
care: knowledge; skills; social/professional role and identity; beliefs and capabilities;
optimism; beliefs about consequences; reinforcement; intentions; goals; memory, attention
and decision processes; environmental context and resources; social influences; emotion;
and behavioral regulation. All domains of the framework will be reviewed for inclusion
during this stage based on their relevance to the research objective.

Refining the questionnaire

Face validity (i.e., extent to which the questionnaire appears to be measuring the
intended information) and content validity (i.e., extent to which questionnaire has
enough coverage of the content area or sufficient number of questions for each domain)
17] will be assessed. At least three primary care physicians, nephrologists, and health
services researchers (n?=?9) will independently review the questionnaire for face and content validity, to
evaluate the merit (e.g., relevance and quality) of each item, to identify unnecessary
or ambiguous questions (to assess clarity, relevance, flow, and wording), and to determine
whether respondents will interpret the questions appropriately and consistently (to
assess interpretation) 18]. The length of time required to complete the questionnaire will also be measured.

Survey administration

Approximately 4500 primary care physicians will be contacted through information publicly
available from the College of Physicians and Surgeons of Alberta 13]. Potential participants will be recruited via mail, fax, and online methods. The
first contact will be a survey package that includes a personalized cover letter with
screening questions to ensure only eligible physicians complete the questionnaire,
a succinct questionnaire, a return fax number and email address, and an online survey
as another option for completing the survey. Based on a modified Dillman approach
19], the first reminder with the same information (via fax) will be sent 1 week, the
second reminder 3 weeks (mail), and the third reminder 4 weeks (fax) after the initial
contact for non-responders.

Quantitative analysis

Questionnaire responses will be reported using descriptive statistics on the proportion
of key barriers and facilitators perceived by primary care physicians in Alberta.
Associations between demographic and practice characteristics and outcome variables
(e.g., self-reported familiarity and use of comprehensive conservative programs) will
be analyzed using regression models with backward elimination techniques. Based on
a sample size calculation, a sample size of 371 eligible respondents will produce
a 95 % confidence interval equal to the sample proportion plus or minus 0.05 when
the estimated proportion is 0.50. At the interpretation stage, the extent to which
the quantitative results extend the initial qualitative findings will be discussed.

Dissemination

Employing an integrated knowledge translation approach, decision-makers (leaders in
primary care as well as the renal programs) have been involved in the development
of study design, including interview and survey questions. Results from this mixed
methods study will be published in peer-reviewed scientific journals, published in
provincial and local newsletters targeted to primary care providers, and presented
at national and international conferences to audiences in the areas of nephrology,
family medicine, and geriatrics. In collaboration with key stakeholders of providers
in the conservative care of patients with stage 5 CKD, the findings of this study
will be used to inform the development of clinical practice tools and resources, including
a provincial conservative kidney management clinical care pathway as a standardized
informational tool to implement clinical practice guidelines and best practices in
renal care, for healthcare providers, patients, and their family members.