Secondary Logo

Journal Logo

Research Reports

Validating Domains of Patient Contextual Factors Essential to Preventing Contextual Errors: A Qualitative Study Conducted at Chicago Area Veterans Health Administration Sites

Binns-Calvey, Amy E.; Malhiot, Alex MS; Kostovich, Carol T. PhD, RN, CHSE; LaVela, Sherri L. PhD, MPH, MBA; Stroupe, Kevin PhD; Gerber, Ben S. MD, MPH; Burkhart, Lisa PhD, RN; Weiner, Saul J. MD; Weaver, Frances M. PhD

Author Information
doi: 10.1097/ACM.0000000000001659

Abstract

Incorporating patient context into clinical decision making is essential to planning effective care. Patient context refers specifically to all that is expressed outside the boundaries of a patient’s skin that is relevant to their care. It can be divided into patient circumstances, such as job loss, and patient characteristics or states, such as progressive dementia. The former is external to the patient (i.e., financial hardship), and the latter internal (cognitive decline). Both could account for deteriorating medication adherence. Each requires a different approach to management. The process of identifying and incorporating these contextual factors into care planning is a highly individualized process called “contextualizing care.” It has been described as answering the question “What is the best next thing for this patient at this time?”1

Inattention to patient context is termed a “contextual error” when it results in a care plan that is inappropriate.2 For instance, increasing the dose of an expensive insulin analogue when a patient’s diabetes control deteriorates, although the patient cannot afford it, would constitute a contextual error. Switching him to less expensive human insulin that he can afford contextualizes his care.3 Hence, contextualizing care and contextual error reflect, respectively, attention vs. inattention, to patient context in clinical decision making and care planning. There is empirical evidence that avoiding contextual errors and contextualizing care predicts both better health care outcomes and reduced overuse and misuse of medical services.4,5

What are the contextual factors in patients’ lives that clinicians must attend to in order to avoid contextual errors?6 No two contextual factors are quite the same. A care plan is optimally contextualized when it is customized to address, to the degree possible, the contextual factors specific to that patient so that she or he has the best chance of having a good outcome.

From the clinician’s perspective, the process of contextualizing care consists of four steps.7 The first is recognizing clues that a contextual factor might be present, a process referred to as recognizing “contextual red flags.”4 Loss of control of diabetes in a patient who was previously well controlled is a contextual red flag because it is a clue that something may have changed in that patient’s life situation that is impacting the patient’s ability to manage his care. The second step is exploring the contextual red flag, referred to as “contextual probing,” typically by asking the patient about the contextual red flag, as in “I notice your diabetes is not well controlled these days; what do you think is going on?” If, in the third step, a contextual factor is identified (e.g., patient with poor eyesight is having difficulty dosing his insulin), the fourth step consists of incorporating that factor into the plan of care (e.g., prescribing prefilled syringes).

Addressing patient contextual factors when present, possible, and practical is associated with improved patient health care outcomes. Weiner and colleagues4 found that patients whose providers addressed their contextual factors demonstrated improved health outcomes (including lower blood pressure, better medication adherence) in 71% of the observed cases, whereas patients whose providers did not incorporate contextual factors into a plan of care demonstrated improved outcomes in only 46% of cases. Failures to contextualize care are also costly.5

Although there are a nearly infinite variety of contextual factors, given the complexities of people’s lives, they can be grouped into broad categories, termed “contextual domains.” Working the night shift (a contextual factor) could be considered a “competing responsibility” (a contextual domain), as could caring for a sick child when doing so precludes receiving necessary care. An inability to pay for a medication (a contextual factor) could be considered a “financial situation” (a contextual domain) when it accounts for a deteriorating clinical condition. A characteristic of all contextual factors, as the term is defined, is that they complicate care planning. They are challenges to address, problems to work around for a particular patient at a particular moment in time. This is distinct from broader terms such as “person-related information” (PeRI) and “contextual information,” which are useful when considering classification systems.8,9 For instance, sociodemographic characteristics are considered a class of PeRI8 but are likely too broad a term to be useful when considering why a patient has stopped taking her medication.

On the basis of clinical experience, Weiner1 initially grouped contextual factors into 10 domains of context: competing responsibilities, social support, access to care, economic situation, skills and abilities, emotional state, cultural perspective, spiritual beliefs, attitude towards illness, and relationship with health care provider. These 10 categories were created as a typology, based on Weiner’s experience in the clinic, and were not subjected to empirical analysis. Although these contextual domains are often referenced as a framework for classifying patient contextual factors, they have not been systematically substantiated by providers or patients.8,10–13 The purpose of this study was therefore to empirically confirm and refine the list of contextual domains using qualitative methods that elicited patient and provider input.

Method

We used qualitative methods to gather data on patient and health care provider perspectives on patient contextual factors that affect health care management and decision making. We conducted semistructured interviews with providers and held focus groups with patients from two Department of Veterans Affairs (VA) medical centers and four community-based outpatient clinics (CBOCs) in the Chicagoland area. CBOCs are health care clinics in separate locations from, but associated with, a larger VA medical facility. One-on-one semistructured interviews that accommodated clinicians’ schedules balanced open inquiry with efficiency. Focus groups with patients allowed us to obtain patient perceptions efficiently and to elicit shared as well as disparate views about the contextual factors in their lives that impact their care.14,15 To increase the diversity of study participants and reflect variation in care delivery, we selected CBOCs to represent rural and urban settings and varied sociodemographic characteristics in addition to VA tertiary care facilities.

The principal investigator (F.W.), project manager (A.B.C.), and research assistant (A.M.), all of whom had prior experience with qualitative data collection, conducted the interviews and focus groups. The transcripts were reviewed and coded by two PhD researchers (F.W., C.K.), a research assistant with experience in patient-centered care research (A.M.), and a team member who has over eight years of research experience in contextualization of health care (A.B.C.).

Provider interviews

We conducted interviews with health care providers at each of the study sites during February and March 2014. All but one interview were conducted via telephone. Interviews lasted between 30 and 55 minutes and were recorded and transcribed. (A recorder failed during one interview, and we used the interviewer’s notes for coding.) We designed questions to elicit the broadest range of contextual factors providers encounter, avoiding leading questions. Questions focused on patient, community, and resource-related factors in their patients’ lives that they have had to take into account when planning care. See List 1 for selected questions. The complete institutional review board (IRB)-approved provider interview guide is online (Supplemental Digital Appendix 1, https://links.lww.com/ACADMED/A436). Note it contains additional items for a separate project to develop a patient-completed inventory on contextual factors.

List 1

Provider Semistructured Interview Questions, From a Study Validating and Identifying Domains of Patient Context, Chicago Area Veterans Health Administration Sites, 2014

Individual

  • When you think about your patients individually, what are some patient-specific issues that emerge within the context of providing care and making decisions (e.g., cognitive abilities, emotional state, attitude toward illness, spiritual beliefs, values and goals)?
  • What about factors that influence behavior change (e.g., diet, exercise, smoking cessation)?
  • What about factors related to their home, work (e.g., travel), or community engagements/responsibilities?

Community

  • What are elements of a patient’s broader community that affect their care (e.g., culture, religion, socioeconomic status)?
  • How do these factors affect patient care?

Resources

  • Are there issues related to access that factor into your ability to care for patients (e.g., finances, resources, transportation, distance to facility)?
  • Health technology is increasingly being incorporated into health care delivery. Currently, home telehealth, clinical video telehealth, secure messaging, mobile applications, interactive voice response, and social media are some ways in which the VA is using technology to deliver care.
  • What types of technologies have you used within the VA?
  • What patient-specific factors influence the use of these technologies?
  • How knowledgeable are your patients about these technologies? How interested in or willing are they to use it?

Patient focus groups

Focus groups were held in conference rooms at the study sites during March–May 2014. Groups ranged in size from 4 to 10 participants. Study team members acted as moderators, and discussions lasted between an hour and an hour and a half. A discussion guide that included open-ended items with probes and follow-up questions was developed; it included questions designed to generate discussion among participants concerning patient contextual factors that affect patients’ abilities to manage their care. These questions also centered on patient-specific, community, and resource-related factors in their lives that they have had to take into account when planning care. See List 2 for selected questions. The complete IRB-approved provider interview guide is online (Supplemental Digital Appendix 2, https://links.lww.com/ACADMED/A478). Note it contains additional items for a separate project to develop a patient-completed inventory on contextual factors. The focus groups were recorded and transcribed.

List 2

Patient Focus Group Discussion Guide, From a Study Validating and Identifying Domains of Patient Context, Chicago Area Veterans Health Administration Sites, 2014

Individual

  • Each patient is unique and has their own preferences for care. What are some things that you would like your doctor to consider when making recommendations and decisions about your health?
  • Please describe any personal beliefs you have about your health that influence how you choose to care for yourself.
  • What role have your personal relationships played in you managing your health (e.g., refilling your prescriptions, cooking)?
  • How can your support system be helpful in managing your health?
  • What is important to you about having others involved in your health outside of you and your doctor?
  • For those of you who have been accompanied by a family member/friend to a health care appointment, describe why you did and how the doctor interacted with this individual.
  • If you could add information to your medical record for your doctors and your health care team to see, what types of things would you include?

Community

  • What are some groups or activities you engage in within your community that may influence your health (e.g., church, social clubs/activities)?
  • How do these things influence your health decisions? (For example, church members sometimes desire to pray about health decisions; support groups provide meals, light chores, etc.)
  • How do issues related to the availability of transportation and safety of your neighborhood influence your health decisions?

Resources

  • Tell me about your experience with technology (e.g., computers, cell phones). What experience do you have using technology to manage your health care?

Participants

We recruited primary care physicians and registered nurses who were either Patient Aligned Care Team (PACT) or primary care team members. A PACT team is an interprofessional team of providers that offers comprehensive treatment to veterans at a VA facility.16 We identified primary care providers by obtaining lists from the facilities and recommendations from facility leadership. From January to March 2014, we sent e-mail invitations to 63 providers. Our target was to interview 3 physicians and 3 nurses each at the two medical facilities and 1 physician and 1 nurse at each of the four CBOC sites. If the target number of participants wasn’t met at a particular facility, a second round of invitations was sent to the nonresponders.

We obtained lists of veterans cared for in primary care clinics at these facilities. To account for site leadership preferences, the protocol for inviting patients varied: For one VA facility and two of its affiliated CBOCs, we were given names for the entire patient panel from these facilities. Using a computer-generated randomized number, we invited every 50th person on the list to participate. For the other VA facility, specific primary care providers gave us their lists of patient panels. Using a computer-generated randomized number, we invited every 10th person on these providers’ panels. At two CBOCs associated with this facility, we were given short lists of patients cared for by participating providers and invited all patients on those lists to participate.

In total, from March to May 2014, 363 patients across the six study sites were sent invitations to participate (with an option to opt out within two weeks). Our goal was to include 40 to 96 veterans in the focus groups. Veterans who did not opt out were contacted by telephone and invited to participate in a focus group meeting.

The study received approval and a waiver of documentation of informed consent process from the VA Central IRB. Patient participants were paid $25 for their time and travel expenses.

Analysis

We transcribed provider interviews verbatim and analyzed them using directed content analysis techniques to code the data thematically. A directed approach to content analysis is appropriate when seeking to validate an extant theory.17 We began using a priori categories based on Weiner’s1 previously identified domains of context as a framework, as listed above. Three research team members independently coded three transcripts and compared results to identify recurring themes and establish an initial code list. Coders agreed to notate any contextual factors that did not fall into the existing framework used for coding. Coders then met to discuss the codes and resolve any discrepancies. Once consensus was reached, the codebook was finalized and the remaining transcripts were distributed among the three research team members to code.

We analyzed patient focus group discussions using the same techniques. Transcripts were distributed among research team members, including a fourth team member, to discuss, reach consensus, and finalize the codes using the identified themes, again with a focus on contextual domains.

Results

Provider participants

Of 63 providers sent e-mail invitations, 19 (30%) verbally consented and were interviewed. One registered nurse and 1 provider at each of three CBOCs were interviewed. A fourth CBOC was short-staffed, so we were only able to interview a nurse from that site. At the two VA medical centers, 3 physicians and 3 registered nurses were interviewed from each site. A total of 19 interviews (18 by telephone, 1 in person) were conducted. Interviews averaged 30 minutes in duration. Provider participants were 84% female.

Patient participants

Recruitment letters were mailed to 363 veterans from the list of 368 patients identified (5 were not sent letters because they were in another study, were no longer patients, or were flagged as being veteran employees/sensitive). Seventy-eight veterans agreed to participate, but 29 either canceled or did not show on the day of the focus group sessions.

Forty-nine veterans (13.5%) participated in one of eight focus groups (two groups at each of the two VA facilities for a total of 22 participants and one focus group at each of the four CBOCs for a total of 27 participants).

Forty-two (86%) participants were male, and 14 (29%) were from rural areas. Twenty-eight (57%) patients volunteered their ages; of those, the age range was 32 to 90 years (mean age 62.9). Twenty-nine (59%) patients also volunteered whether they received their care solely from the VA or had another non-VA provider. Of those who provided that information, 16 (55%) saw VA providers solely.

Assessment of domains of context

Coders noted when participants discussed patient contextual factors that fell within 1 of the 10 previously identified domains. Each of the original domains of context was mentioned during the provider interviews and patient focus group discussions. However, through analysis of the data, coders identified additional contextual factors mentioned by participants that could not be classified into 1 of the 10 previously identified domains. On the basis of comments made by both patients and providers, we thus made modifications to the original contextual domains to better represent responses. Four of the original domains were modified, and 3 new domains were identified. Coders developed a revised list of domains based on the themes that emerged from the data, resulting in 12 domains (see List 3).

List 3

Domains of Patient Context: 6 Circumstances and 6 Characteristics/States, From a Study Validating and Identifying Domains of Patient Context, Chicago Area Veterans Health Administration Sites, 2014

Circumstances

  • Access to care: The patient’s ability to receive care in a timely manner.
  • Competing responsibility: An obligation or commitment the patient has that impacts their ability to manage their health care.
  • Social support: A patient’s access to a supportive network of individual(s) able to assist if needed.
  • Financial situationa: The patient’s ability to afford care.
  • Environmentb: The physical and social setting that encompasses a patient.
  • Resourcesb: The possessions and materials available to a patient that can facilitate a person’s ability to manage their care.

Characteristics/States

  • Skills, abilities, and knowledgea: A patient’s intellectual understanding and physical ability to manage health care.
  • Emotional state: The emotional condition of a patient as it relates to their ability to manage their health care.
  • Cultural perspective/spiritual beliefsa: The customs or a faith-based practice a patient has that impact health care.
  • Attitude towards illness: The feelings a patient has towards their condition that impact their ability to manage it.
  • Attitude towards health care provider and systema: The patient’s feelings and attitudes towards their providers and the health care system that impact their ability to manage their health care.
  • Health behaviorb: The patient’s actions and lifestyle choices that impact their health care.

aModified from Weiner SJ. Contextualizing medical decisions to individualize care: Lessons from the qualitative sciences. J Gen Intern Med. 2004;19:281–285.

bNew domain identified by this study.

Original domains that were retained on the basis of contextual factors noted by providers and patients were the following:

  • Access to care is the patient’s ability to receive care in a timely manner. Examples: the distance a patient needs to travel to a health care facility, hours the clinic is open.
  • Competing responsibility is an obligation or commitment the patient has that impacts her ability to manage her health care. Examples: caretaker responsibilities, job requirements.
  • Social support is a patient’s access to a supportive network of individual(s) able to assist if needed. Examples: a spouse who helps with medications, a neighbor who drives the patient to appointments.
  • Financial situation is the patient’s ability to afford care. Examples: cost of medications; ability to afford transportation to clinic.
  • Skills, abilities, and knowledge include a patient’s intellectual understanding and physical ability to manage health care. Examples: physical ability to take a medication (read the dosage on a syringe); understanding of how to schedule appointments; understanding provider instructions or the purpose of medications prescribed.
  • Emotional state refers to the emotional condition of a patient as it relates to his ability to manage his health care. Examples: incapacitating depression, fear of procedures.
  • Cultural perspective/spiritual beliefs comprise the customs or a faith-based practice a patient has that impacts health care. Examples: a culture that emphasizes herbal remedies, a religious faith that discourages certain types of medications, a reliance on faith instead of medical advice.
  • Attitude towards illness includes the feelings a patient has towards his or her condition that impact the patient’s ability to manage it. Example: stigma of having a condition.
  • Attitude towards health care provider and system is focused on the patient’s feelings and attitudes towards his or her providers and the health care system that impact the patient’s ability to manage his or her health care. Examples: contentious relationship between patient and provider, distrust of medical facility.

In addition to affirming the aforementioned domains, patient and provider participants offered insights that prompted us to refine and clarify, where needed, some of the domains. Three of the previously identified domains were modified on the basis of the data from the interviews and focus groups.

The previous domain “economic situation” was reworded to “financial situation.” On the basis of feedback from participants, we concluded that financial situation was a more precise characterization of the patient’s individual circumstances; economic situation could denote the general state of the economy on a larger scale.

The previous domain, “skills and abilities,” was modified to include “knowledge.” While a patient’s knowledge of his or her condition and treatment was previously included in the domain of “skills and abilities,” we concluded that adding the term “knowledge” more accurately described this domain.

In the previous list, “cultural perspective” and “spiritual beliefs” were separate categories. We combined the two, as they were viewed similarly by our respondents, and it could be argued that a spiritual belief reflects a cultural perspective.

We renamed the domain “relationship with health care provider” to “attitude towards health care provider and system.” We replaced the word “relationship” with “attitude” because attitude drives behavior. The term “relationship” introduces characteristics of another party other than the individual patient. We added “and system” because patients’ attitudes also shape their interactions with the health care system as a whole.

Finally, we introduced the following three new domains identified during interviews and focus groups in response to hearing contextual factors that did not fit appropriately within those already described:

  • Environment is the physical and social setting that encompasses a patient. Examples: patient lives in unsafe neighborhood where medications are stolen; patient has a social group that engages in unsafe behavior.
  • Resources are the possessions and materials available to a patient that can facilitate a person’s ability to manage his or her care. Examples: computers, cell phones, glucose meters, blood pressure monitors, adaptive equipment, and mobility aids.
  • Health behavior describes the patient’s actions and lifestyle choices that impact his or her health care. Example: engaging in unsafe or reckless activities even when acknowledging the risks would be a health behavior.

Of the 12 domains explicated, we noticed that 6 domains encompass circumstantial contextual factors (e.g., job loss, not having transportation). The circumstance domains are access to care; competing responsibilities; social support; financial situation; environment; and resources. The other 6 domains encompass contextual factors consisting of internal patient characteristics or states at the time of the encounter—emotional state; cultural/spiritual; skills, abilities and knowledge; attitudes toward illness; attitudes toward providers and health system; and health behavior—that are manifest in particular behaviors. Table 1 contains a sampling of participants’ statements for each of the identified domains of context.

T1
Table 1:
Participant Statements for 12 Identified Domains, From a Study Validating and Identifying Domains of Patient Context, Chicago Area Veterans Health Administration Sites, 2014

Discussion

Effective care planning requires identifying patient-specific contextual factors that, if unaddressed, would undermine the goals of care. Although there are innumerable such factors, reflective of the infinite variation in the unique life circumstances of every patient as an individual, according to our findings they cluster into 12 broad domains. The domains can serve as a “differential” for considering the possible reasons why a patient’s care may not be proceeding as planned, or requires redesign. Identifying domains of context can help providers formulate an appropriate plan of care, thereby avoiding contextual error.

The revised 12 domains of context that coders identified may serve as a cognitive tool to facilitate provider recognition and incorporation of patient context into medical decision making. As noted, these domains fall into two groups: The first 6 are circumstantial, encompassing everything that is literally “outside the patient’s skin” as noted earlier (e.g., financial situation, competing responsibilities). In contrast, domains 7 through 12 are patient characteristics/states, such as an individual’s skills, cultural perspective, and emotional state, that are expressed outside the boundaries of a patient’s skin as behaviors—such as when a depressed patient loses control of her diabetes because she stops taking her medication. Depression becomes a contextual factor essential to address when planning her diabetes care.

Providers may be more attuned to issues in patients’ lives once they familiarize themselves with the domains into which a spectrum of contextual factors falls. Medical and nursing education could incorporate the list of domains as areas for students to consider in all encounters when there are clues (i.e., contextual red flags) that point to a contextual dimension of care.

Limitations

The categories we have identified were based on our data collection with veteran primary care patients and providers of primary care. It is possible that there are unidentified contextual factors relevant to care delivery that would not fit into 1 of the 12 identified domains. Examination of these domains in the context of other nonveteran and non-VA settings is warranted.

Conclusions

To tailor medical care to the needs of each patient at a specific time in his or her life, information on patient contextual factors must be integrated with evidence-based practice in health care planning. When context is taken into consideration, health care outcomes improve, costs may be reduced, and care is free of error resulting from inattention to patient context. A wide array of contextual factors that impact patients’ ability to manage their health and health care can be grouped into 12 domains, providing a framework for thinking about patient context in care planning.

References

1. Weiner SJ. Contextualizing medical decisions to individualize care: Lessons from the qualitative sciences. J Gen Intern Med. 2004;19:281285.
2. Weiner SJ. Kattan M. Contextual error. In: Encyclopedia of Medical Decision Making. 2009:Thousand Oaks, CA: Sage; 198202.
3. Tylee T, Hirsch IB. Costs associated with using different insulin preparations. JAMA. 2015;314:665666.
4. Weiner SJ, Schwartz A, Sharma G, et al. Patient-centered decision making and health care outcomes: An observational study. Ann Intern Med. 2013;158:573579.
5. Schwartz A, Weiner SJ, Weaver F, et al. Uncharted territory: Measuring costs of diagnostic errors outside the medical record. BMJ Qual Saf. 2012;21:918924.
6. Weiner SJ, Kelly B, Ashley N, et al. Content coding for contextualization of care: Evaluating physician performance at patient-centered decision making. Med Decis Making. 2014;34:97106.
7. Weiner SJ, Schwartz A. Contextual errors in medical decision making: Overlooked and understudied. Acad Med. 2016;91:657662.
8. Schrans D, Avonts D, Christiaens T, et al. The search for person-related information in general practice: A qualitative study. Fam Pract. 2016;33:9599.
9. Weir CR, Staggers N, Gibson B, Doing-Harris K, Barrus R, Dunlea R. A qualitative evaluation of the crucial attributes of contextual information necessary in EHR design to support patient-centered medical home care. BMC Med Inform Decis Mak. 2015;15:30.
10. Col NF. Challenges in translating research into practice. J Womens Health (Larchmt). 2005;14:8795.
11. Dunlea R, Lenert L. Understanding patients’ preferences for referrals to specialists for an asymptomatic condition. Med Decis Making. 2015;35:pii: 0272989X14566640.
12. Mount JK, Massanari RM, Teachman J. Patient care complexity as perceived by primary care physicians. Fam Syst Health. 2015;33:137145.
13. Thompson BM, Teal CR, Scott SM, et al. Following the clues: Teaching medical students to explore patients’ contexts. Patient Educ Couns. 2010;80:345350.
14. Krueger RA. Focus Groups: A Practical Guide for Applied Research. 1998.Newbury Park, CA: Sage Publications.
15. Morgan DL. Focus Groups as Qualitative Research. 1997.2nd ed. Thousand Oaks, CA: Sage Publications.
16. Kearney LK, Post EP, Pomerantz AS, Zeiss AM. Applying the interprofessional patient aligned care team in the Department of Veterans Affairs: Transforming primary care. Am Psychol. 2014;69:399408.
17. Hsieh HF, Shannon SE. Three approaches to qualitative content analysis. Qual Health Res. 2005;15:12771288.

Supplemental Digital Content

Copyright © 2017 by the Association of American Medical Colleges