Page 5 - Kansas Journal of Medicine, Volume 10 Issue 1
P. 5

KANSAS JOURNAL of M E D I C I N E                                    ship fees, reported as continuous variables, were aggregated
                                                                     into interval ranges allowing comparison to the prior studies.
DIRECT PRIMARY CARE                                                  Data are presented as a percentage of respondents reporting.
continued.
                                                                     RESULTS
METHODS                                                              2015 Survey Results
	 Participants. This study was deemed “non-human subjects”           	 Practice demographics, physician salaries, and work hours.
research by the Institutional Review Board at the University of      Of the 65 direct primary practices sent an invitation to partici-
Kansas School of Medicine-Wichita. Potential practices to sur-       pate in the survey, responses were received from 38 (59% re-
vey were identified using the Googleâ„¢ search engine for the          sponse rate), representing 20 different states (Figure 1). The
terms “direct primary care” and “concierge medicine.” Prac-          majority of respondents (74%) reported physicians spend fewer
tices that clearly self-identified with these labels were chosen     than 50 hours per week devoted to patient care and practice
from the top 100 search results. If no practice email was read-      management (Table 1). Most (72%) reported projected physi-
ily available on the website, a phone call was made to request       cian incomes between $200,000 and $300,000, and half report-
participation and contact information. Additional participants       ed current physician incomes between $100,000 and $200,000
were identified using snowball sampling, wherein respon-             (Figure 2). Two-thirds (65%) of respondents reported their
dents suggested other DPC practices to receive the survey.           practice as urban, and one-third (35%) as rural. Sixty-one per-
	 Instruments. Data were collected using a survey instrument         cent (61%) of respondents reported having transitioned from
distributed via the Survey Monkey® online platform. The sur-         a traditional practice and 39% were de novo practices. Most
vey instrument was developed solely for this study, based on         (70%) respondents reported being in practice for one year
the intent to compare data to previous characterizations. Ad-        or less, and reported having fewer than six employees (94%).
ditional items were added to obtain practice demographics, fi-       	 Practice amenities are presented in Figure 3. The majority
nancial characteristics, and to expand on the possible amenities     of respondents (84%) reported using an EMR, offering physi-
offered as described below. The continuous variables included        cian email access (82%), 24-hour access (76%), same day ap-
years in practice, the number of physicians in the practice, num-    pointments (92%), and wholesale labs (74%). Few respon-
ber of staff members, and membership fees. Physician salary,         dents offered inpatient care (16%), obstetrics (3%), or financial/
work hours, and patient panel size were collected using inter-       insurance consultant services (18%). Survey data regarding
val values. Physician salary and patient panel size were col-        these points were not available from 2005 for full comparison.
lected in current and projected forms, with ‘projected’ being        Selected Comparisons to 2005
defined as the desired end point for the practice, rather than a     	 Specialties, membership fees, insurance billing, and projected
distinct time period. Discrete variables included state of prac-     panel size were compared between 2005 and 2015. The results
tice, practice setting (rural or urban), residency training, accep-  are presented in Figure 4. In 2015, most respondents (87%) were
tance of Medicare patients, size of practice, and the presence or    family medicine physicians; 5% were internal medicine physi-
absence of numerous services and amenities including: elec-          cians, with the remainder from pediatrics or internal medicine/
tronic medical record (EMR), patient portal, physician email         pediatrics. This was a statistically significant reduction from the
access, social networking (i.e., Twitter, Facebook), financial or    62 - 77% of DPC practices reported as internal medicine training
insurance patient consultant, 24-hour physician access, same         in 2005 (p < 0.0001, compared to the 62% to be conservative).
day appointments, house calls, inpatient care, obstetric care,       Second, the majority (88%) of respondents reported annual
wholesale labs, wholesale medications, wholesale imaging,            individual adult membership rates of $500 - $1,499, a signifi-
employer group contracts, and immunizations. Additionally,           cant increase from 2005 where 19% reported between $500 and
there were two free text fields for any additional comments, as      $1,499 (p < 0.0001). Third, few practices (11%) reported submit-
well as for referral contact information for other DPC practices.    ting bills to insurance in 2015, a statistically significant decrease
	 Procedures. Sixty-five practices were identified. They were        from the 75% in 2005 (p < 0.0001). Last, fifty-six percent (56%)
sent an email containing a link to the online survey that re-        of practices reported projected patient panel size ≥ 600 in 2015,
quested their participation in the study. In the initial email,      a statically significant increase from 40% in 2005 (p = 0.0274).
they were assured of anonymous data reporting. A remind-
er email was sent four weeks later. There was a subsequent           DISCUSSION
four-week interval until data were collected for analysis.           	 DPC, once known as a model focused on providing care for
	 Analysis. Data were collected and analyzed using the SAS           the wealthy, appears to be undergoing a transformation into
software for Windows version 9.3 (Cary, NC). Descriptive sta-        a model that is more accessible to the general population. Al-
tistics were presented as frequencies and proportions for cat-       though 58% of the practices surveyed reported current physi-
egorical variables. A one-sided binomial proportion compari-         cian patient panel sizes of fewer than 400 patients (Figure 2), a
son was conducted using PROC FREQ. The 2015 data were                large number of this study’s respondents reported being in prac-
compared to the corresponding 2005 proportions. Member-              tice for one year or less, and the intention to grow was present.

                                                                                                                                           4
   1   2   3   4   5   6   7   8   9   10