Patient Dismissal by Primary Care Practices

Abstract

This survey study examines the prevalence of and reasons for patient dismissal among primary care practices participating in the Comprehensive Primary Care initiative and a matched sample of comparison practices.

Patient dismissal, the termination of a relationship with a patient by a health care professional, can affect access to care. Although understudied, the topic is important, especially as insurers begin to shift payment from volume to value. Patient dismissal could be an unintended consequence of this shift as clinicians face (or perceive that they face) pressure to limit their panel to patients for whom they can readily demonstrate value in order to maximize revenue. We examined the prevalence of and reasons for patient dismissal among primary care practices participating in the Comprehensive Primary Care (CPC) initiative and a matched sample of comparison practices. The CPC and responding comparison practices had similar market-, practice- and patient-level characteristics.

Methods

The New England Institutional Review Board (NEIRB 13-174) exempted the study based on the federal common rule (section 45 CFR 46.101[b][5]), because the evaluation’s purpose is to evaluate a public benefit program. We analyzed the practices’ responses to a survey fielded in 2016, the last year of the 4-year CPC initiative. The survey went to practice managers who were asked to discuss their responses with others in the practice. All 443 CPC practices and 42% (351) of 849 comparison practices responded, yielding 794 practices. Participating CPC practices were not compensated. Comparison practices and withdrawn CPC practices were compensated up to $125 for responding.

In addition to questions on the practices’ characteristics and approaches to delivering care, the survey asked: “In the past 2 years, has your practice ever dismissed a patient from your practice? By dismissing patients, we mean directing patients to leave your practice and seek primary care elsewhere.” Respondents who answered “yes” were then asked to identify, from specified response categories, reasons for patient dismissal (Table). They were also asked to estimate the number of patients dismissed in the past 2 years by choosing any of the following ranges: 1 to 5, 6 to 10, 11 to 20, 21 to 50, or 51 or more. CPC practices were also asked whether participating in the initiative made it more or less likely that they would dismiss patients. We present descriptive statistics from these data on patient dismissal for CPC practices, comparison practices, and overall.

Table. Distribution of Practices’ Dismissal of Patients and Reasons for Dismissal.

Characteristic	No. (%)a			P Valueb (CPC vs Comparison)
	All Practices	CPC Practices	Comparison Practices
No. of practices in sample	794	443	351
Practice dismissed patients in the past 2 y	708 (90.5)	394 (88.8)	314 (91.9)	.14
For practices that reported dismissing patients in the past 2 y, reasons for dismissalc
Patient repeatedly missed appointments	504 (74.4)	289 (73.1)	215 (75.4)	.48
Patient repeatedly violated bill payment policies	263 (39.5)	139 (35.0)	124 (43.1)	.03
Patient violated chronic pain/controlled substance policies	552 (78.4)	310 (78.7)	242 (78.2)	.86
Patient was extremely disruptive and/or behaved inappropriately toward clinicians or staff	567 (81.2)	323 (81.8)	244 (80.8)	.73
Patient repeatedly did not follow health care recommendations (eg, medication regimens or completing laboratory tests)	313 (44.7)	166 (42.1)	147 (46.8)	.20
Patient repeatedly did not follow recommended lifestyle changes (eg, diet, exercise, or smoking cessation)	38 (6.8)	20 (5.0)	18 (8.3)	.09
Patient made frequent visits to the emergency department and/or frequently self-referred to specialists	40 (5.9)	27 (6.8)	13 (5.2)	.36
Other	21 (3.3)	18 (4.4)	3 (2.4)	.15
Distribution of practices by number of patients dismissed in the past 2 years
0	86 (9.6)	49 (11.2)	37 (8.2)	.11
1-5	223 (26.7)	120 (27.5)	103 (26.0)
6-10	170 (23.9)	88 (19.9)	82 (27.2)
11-20	137 (16.6)	81 (18.1)	56 (15.3)
21-50	103 (14.9)	65 (14.8)	38 (15.1)
≥51 patients	70 (8.4)	38 (8.6)	32 (8.3)
Participation in CPC made practice more or less likely to dismiss patients
Much more likely	NA	3 (0.8)	NA	NA
Somewhat more likely	NA	17 (4.2)	NA
Neither more or less likely	NA	324 (82.8)	NA
Somewhat less likely	NA	30 (7.5)	NA
Much less likely	NA	19 (4.8)	NA
If participation in CPC made practice more likely to dismiss patients, reasons whyd
Concerns about meeting quality metrics for CPC	NA	16 (79.5)	NA	NA
Concerns about meeting financial metrics for CPC	NA	5 (25.6)	NA
Other	NA	4 (20.5)	NA

Abbreviations: CPC, Comprehensive Primary Care initiative practices; NA, not applicable because the question was only asked to CPC practices.

^aFrequencies do not always add to the total number of practices owing to small numbers of missing responses for particular items. Percentages are weighted for nonresponse and matching and therefore may differ from the frequencies.

^bWe used 2-tailed t tests and χ² tests to assess the difference between CPC and comparison practices.

^cPractices could mark all reasons that applied.

^dDenominator is the 5% of practices that indicated that CPC participation made them more likely to dismiss patients. Practices could mark all reasons that applied.

Results

A similar proportion and distribution of CPC and comparison practices reported ever dismissing patients in the past 2 years (394 [89%] and 314 [92%], respectively). Most practices reported dismissing only a few patients in the past 2 years; 86 (about 10%) did not dismiss any patients, and 530 (67%) reported dismissing 1 to 20 patients (Table). The number of patients dismissed was proportional to practice size (data not shown). CPC and comparison practices dismissed patients for similar reasons. The exception was that comparison practices more frequently reported dismissing patients for violating bill payment policies than did CPC practices (124 [43%] vs 139 [35%]).

Practices that reported dismissing any patients did so for the following reasons: the patient was extremely disruptive or behaved inappropriately toward clinicians or staff (567 [81% of practices]), violated chronic pain and controlled substance policies (552 [78%]), or repeatedly missed appointments (504 [74%]) (Figure). Fewer practices reported dismissing patients for repeatedly not following medical recommendations (313 [45%]), violating bill payment policies (263 [39%]), repeatedly not following recommended lifestyle changes (38 [7%]), and making frequent visits to the emergency department or self-referring to specialists (40 [6%]).

Figure. Practices’ Reported Reasons for Patient Dismissal, Among 708 Practices That Dismissed Any Patients.

Data for reported reasons are from the Comprehensive Primary Care (CPC) initiative and comparison practices combined. Practices could mark multiple responses. Data were collected in the 2016 CPC Practice Survey.

Discussion

Our results provide early evidence on the influence of one alternative payment model on patient dismissal and the reasons for it. According to most CPC practices, the initiative had no effect or made them less likely to dismiss patients. The CPC and comparison groups dismissed patients for similar reasons.

The limitations of our study include the use of survey data, which may be subject to recall or social desirability biases. We used survey data because it was not feasible to review patient medical records or dismissal letters from the 1292 practices.

Our study contributes new insight into an issue that will be increasingly important as insurers move to reimbursing for value rather than volume. Future research could investigate what medical ethicists consider inappropriate reasons for dismissal.