Case Management in Primary Care for Frequent Users of Health Care Services: A Mixed Methods Study

The purpose of this study was to evaluate the effect of the Vulnerable Patients in Primary Care: Nurse Case Management and Self- management Support (V1SAGES) intervention in a primary care setting for patients with multiple comorbidities. The authors of this study propose that appropriately trained case managers decrease patients' psychological distress and motivate them to take control of their health. Although the outcomes of the study are clearly stated, the term "patient activation" is not clearly defined at any point in the article. We interpret the authors to mean patients will have the knowledge and skillset to effectively manage their health and healthcare.

The study utilized a mixed-methods approach of qualitative and quantitative assessment; qualitative methods assess the context of a research question and quantitative methods assess the impact. It is not clear where the recruitment of participants took place, but identification started with frequent users of hospital services, which was defined as three of more ED visits or hospitalizations in the past year or heavy clinic users. Family physicians then identified individuals from the initial screening or frequent visitors to their clinic who they believed would benefit from case management intervention. Although there were eligibility criteria (age range and diagnosis of one or more chronic diseases), there is a selection bias in having physicians utilize their judgment to select long-standing patients to be entered into the study. This patient selection process also undermines the strength of the study as a double-blinded, randomized control trial.

The intervention group had the help of nurse case managers, who received "50 hours of theoretical and practical training on case management," as well as four required components of management: evaluation of healthcare needs, personally tailored service plan involving healthcare and community partners, coordination between the partners, and self- management support for patients and their families. In addition, the patients and their spouses were invited to weekly peer-led meetings on how to improve their physical and mental well-being. It is difficult to determine the treatment fidelity as the methods only looked into making sure the four main components were delivered, with no further details. The case management intervention seems more psychologically focused than what is generally seen in the primary care clinics. It is not clear whether the peer-leaders were involved with the V1SAGES case management intervention or a separate service offered to the patients in the intervention group, which would be an important distinction to make as the outcomes reported are stated to be a result of the intervention. The control group received the "usual care" from primary care nurses, including health promotion, global evaluation, and follow-up of chronic diseases. It is not clear the amount of interaction the intervention or the control group had with their respective nurses. This information would help determine whether any benefits from case management are due to their specialized training and program or the quality and standardization of "usual care." The article also states that each case manager could have up to 50 patients, which seems like a daunting task and may limit the efficacy of their intervention. A final count of the average number of patients per case manager would help put the impact of the outcomes into perspective of the time that was allocated to each patient.

The quantitative assessment demonstrated a significant difference in psychological distress between the intervention and control group, with the intervention group showing a greater decrease in psychological distress after the six-month period. There was no difference in patient activation between the two groups. Details of the quantitative data analysis have one major omission: there is no explanation for the variables used in the logistic regression analysis to determine the adjusted odds ratio presented in Table 2. The article states that the regression was used to adjust "for baseline presence of psychological distress. No other variable was included in this analysis because there were no stratification variables to take into account." This is not a complete analysis since the power of regression is being able to take into account the different variables. There may have been no need to perform a regression at all and instead a simple odds ratio could have been reported. The analysis also included other instruments that were not reported on, such as disease burden and morbidity assessment. A stress assessment, with a tool such as the Perceived Stress Scale[2], should have been implemented to decrease the likelihood of confounding variables in the two groups.

The qualitative component utilized semi-structured focus groups, whose conversations were recorded, transcribed verbatim, qualitatively coded, and finally processed through triangulation. This component of the article was likely time- and resource-consuming, but provided an additional dimension to the final analysis of a complex interaction between patients and healthcare services. Triangulation showed convergence in psychological distress, but not patient activation. This is an interesting finding as patients reported they felt more empowered in the qualitative analysis. As the article mentions, this may be due to the inadequacy of the Patient Activation Measure.

The article has many strengths such as being a double blind, randomized control trial, and utilizing the aforementioned mixed methods approach to provide further depth and context for the research question. The article also has quite a few limitations, including some of the points mentioned by the authors. The role of the case managers and what exactly was accomplished with their interactions with patients needs clarification. This is important because intervention patients may be more likely to report better care due to the fact they were having more interaction and consistent care compared to the control group, regardless of case management specific interventions. The quality measure should also include patients from the control group to determine if they subjectively experienced a decrease in psychological stress, possibly from social support or acclimation to the health system. There may be some selection bias as the physicians were not blinded to which of their patients were in the study, in addition to the cross-over effect, which is the result of patients in the intervention group interacting with patients from the control group. No attempts to minimize this were noted; intervention patients could have been sampled from three clinics separate from the clinics sampling controls. It is also important to recognize that the twenty-four patients from the intervention group who had been in the Chronic Disease Self-Management Intervention prior to the study constitute a large portion of the control group. This difference in baseline characteristics has significant potential to have an impact on the outcomes.

Given additional time and resources, this study may have benefitted from a longer follow-up time period, blinding physicians to limit the effects of their established relationship with the patient, and stratifying the intervention patients into mild, moderate, and severe groups. The latter requires a significant increase in the study sample size, which may be difficult to accomplish using local hospitals and clinics.

Some modifications to the intervention methods described in the paper may improve the impact of case management in different clinic settings: build the case management into a patient's appointment with their physician to limit multiple trips and related expenses; have case management also address basic patient needs such as housing, food, and transportation; and train individuals other than nurses in case management to decrease the case load for each manager. It may be beneficial to encourage teamwork to triage what should be addressed with the physician versus the case managers. Lastly, develop a model of social determinants of health and how to address them more effectively to affect clinical conditions. The next step of this study should attempt to address the clinical outcomes of reducing stress in patients with multiple comorbidities.

References:

1. Hudon, Catherine, et al. "Case Management in Primary Care for Frequent Users of Health Care Services: A Mixed Methods Study." The Annals of Family Medicine, vol. 16, no. 3, May 2018, pp. 232-239., doi:10.1370/afm.2233.

2. Cohen, Sheldon, et al. "A Global Measure of Perceived Stress." Journal of Health and Social Behavior, vol. 24, no. 4, 1983, pp. 385-396. JSTOR, JSTOR, www.jstor.org/stable/2136404.

Competing interests: None declared