Patient-Reported Social Risks and Clinician Decision Making: Results of a Clinician Survey in Primary Care Community Health Centers

Arwen Bunce; Jenna Donovan; Megan Hoopes; Laura M. Gottlieb; Molly Krancari; Rachel Gold

doi:10.1370/afm.2953

Article Figures & Data

Tables

Additional Files

View popup

Table 1.

Clinician and Survey Summary (N = 38)

Characteristic	Data
Credentials, No. (%)
Physician	9 (23.7)
Advanced practice provider^{^a}	29 (76.3)
Sex, No. (%)
Female	33 (86.8)
Male	5 (13.2)
Number of surveys completed, mean (SD)	16 (9.1)
Ever used SDOH to inform care in surveyed encounters, No. (%)	30 (78.9)
Always used SDOH to inform care in surveyed encounters, No. (%)	2 (5.3)
Never used SDOH to inform care in surveyed encounters, No. (%)	8 (21.1)

SDOH = social determinants of health.
Note: Clinicians came from 21 clinics nested within 14 health systems. Most clinics had 2 clinicians participate in the card study; 2 clinics only recruited 1 clinician and 2 clinics had usable data from only 1 clinician.
↵a Nurse practitioner or physician assistant; 1 recruited clinician was a registered nurse.

View popup

Table 2.

Clinician, Encounter, and Patient Characteristics of Card Surveys

Characteristics	Total Surveys, No. (%) (N = 610)	SDOH Influenced Care, No. (%) (n = 212)	SDOH Did Not Influence Care, No. (%) (n = 398)
Clinicians
Credentials
Physician	127 (20.8)	28 (13.2)	99 (24.9)
Advanced practice provider^{^a}	483 (79.2)	184 (86.8)	299 (75.1)
Sex
Female	547 (89.7)	197 (92.9)	350 (87.9)
Male	63 (10.3)	15 (7.1)	48 (12.1)
Encounters
Type
In-person	504 (82.6)	179 (84.4)	325 (81.7)
Telehealth	106 (17.4)	33 (15.6)	73 (18.3)
Reason for visit
Acute care	152 (24.9)	53 (25.0)	99 (24.9)
Behavioral health	15 (2.5)	3 (1.4)	12 (3.0)
Chronic condition	113 (18.5)	33 (15.6)	80 (20.1)
Hospital follow-up	21 (3.4)	13 (6.1)	8 (2.0)
Routine care	309 (50.7)	110 (51.9)	199 (50.0)
Insurance type
Medicaid	285 (46.7)	101 (47.6)	184 (46.2)
Medicare	132 (21.6)	33 (15.6)	99 (24.9)
Private	107 (17.5)	30 (14.2)	77 (19.3)
Uninsured	86 (14.1)	48 (22.6)	38 (9.5)
SDOH in EHR at encounter?
Yes	166 (27.2)	85 (40.1)	81 (20.4)
No	444 (72.8)	127 (59.9)	317 (79.6)
Pre/post–COVID-19 pandemic^{^b}
Pre	318 (52.1)	135 (63.7)	183 (46.0)
Post	292 (47.9)	77 (36.3)	215 (54.0)
Patients
Sex
Female	353 (57.9)	109 (51.4)	244 (61.3)
Male	257 (42.1)	103 (48.6)	154 (38.7)
Age, y
0-10	14 (2.3)	10 (4.7)	4 (1.0)
11-18	23 (3.8)	8 (3.8)	15 (3.8)
19-29	79 (13.0)	23 (10.8)	56 (14.1)
30-39	102 (16.7)	29 (13.7)	73 (18.3)
40-49	108 (17.7)	46 (21.7)	62 (15.6)
50-59	131 (21.5)	52 (24.5)	79 (19.8)
≥60	153 (25.1)	44 (20.8)	109 (27.4)
Race/ethnicity
Hispanic/Latine	131 (21.5)	60 (28.3)	71 (17.8)
Non-Hispanic White	300 (49.2)	71 (33.5)	229 (57.5)
Non-Hispanic Black	82 (13.4)	38 (17.9)	44 (11.1)
Non-Hispanic Asian	59 (9.7)	32 (15.1)	27 (6.8)
Other^{^c}	38 (6.2)	11 (5.2)	27 (6.8)
Preferred language
English	457 (74.9)	124 (58.5)	333 (83.7)
Spanish	96 (15.7)	52 (24.5)	44 (11.1)
Nepali	26 (4.3)	22 (10.4)	4 (1.0)
Chinese	13 (2.1)	2 (0.9)	11 (2.8)
All others	18 (3.0)	12 (5.7)	6 (1.5)

EHR = electronic health record; SDOH = social determinants of health.
↵a Nurse practitioner or physician assistant; 1 recruited clinician was a registered nurse.
↵b Pre–COVID-19 pandemic for surveys completed before March 2020, post–COVID-19 for surveys completed after March 2020. No surveys completed in March 2020.
↵c Other includes multiple races, American Indian, Alaska Native, or missing data.

View popup

Table 3.

Card Survey Questions and Responses

Question and Response Options for All Surveys (N = 610)	No. (%)^{^a}
1. Which of the following factors influenced the care you provided during this patient encounter? (Mark all that apply)
Clinical factors (eg, history, laboratory results, medications, etc.)	548 (89.8)
Patient-specific social and economic conditions (SDOH)	212 (34.8)^{^b}
Patient preferences	450 (73.8)
Available resources (eg, availability of specialty care, formulary restrictions, insurance requirements)	334 (54.8)
Other (free text):^{^c}	57 (9.3)
Language	28
Education/(health) literacy/knowledge/understanding	15
Transportation	15
Financial/cost/(lack of) health insurance	9
New patient	5
Motivation	4

Questions and Response Options for Surveys With Care Influenced by SDOH (n = 212)	No. (%)^{^a,^d}
2. How did you know this patient-specific SDOH information? (Mark all that apply)
Prior personal knowledge of the patient	135 (63.7)
Conversation with the patient during this encounter	161 (75.9)
Communication from other clinic staff	38 (17.9)
Review on paper SDOH screen	25 (11.9)
Reviewed in EHR	97 (45.8)^{^e}
Other	7 (3.3)
3. The amount of time it took to look up the patient-specific SDOH information was:
Just right, didn’t take long at all	151 (71.2)
Longer than I expected	5 (2.4)
Too long, I gave up	1 (0.5)
N/A, I didn’t try	46 (21.7)
Other	9 (4.2)
4. Did you wish you had SDOH information that was not available?
Yes	29 (13.7)
No	182 (85.8)
Missing	1 (0.5)
5. How much did the patient’s SDOH information influence the clinical decisions (eg, treatment decisions, medications prescribed) you made?
A lot	73 (34.4)
Somewhat	101 (47.6)
Not at all	38 (17.9)
6. How much did the patient’s SDOH information influence any additional care you provided (eg, social service or behavioral health referrals, warm hand-offs)?
A lot	54 (25.5)
Somewhat	97 (45.8)
Not at all	49 (23.1)
Not applicable	12 (5.7)
7. Optional: How did your care change?^{^f}	26 (12.3)
Referred to external low/no-cost programs for clinical care (eg, Imaging for a Cause, GoodRx)	8
Identified need for case management or home health	4
Adjusted follow-up schedule	4
Adjusted medication orders due to cost/lack of insurance	3
Considered mental health as potential source of symptoms or complicating factor	3
Unable to help because the clinic doesn’t have necessary resources	2
Called landlord regarding living conditions	1
Connected patient with food assistance	1

EHR = electronic health record; SDOH = social determinants of health.
↵a Some questions allowed multiple responses, thus numbers and percentages may not sum to total.
↵b Of these surveys, 85 (40.1%) had standardized SDOH screening data in the patient’s EHR at time of encounter.
↵c Each survey could have multiple free-text responses. Responses appearing in >3 surveys are reported and grouped thematically by authors.
↵d Questions 2-6 were only asked on surveys that selected “patient-specific social and economic conditions (social determinants of health)” as a response to question 1.
↵e Of these surveys, 45 (46.4%) had standardized SDOH screening data in the patient’s EHR at time of visit.
↵f Multiple responses allowed. All responses reported and grouped thematically by authors.

View popup

Table 4.

Odds Of Card Survey Response Indicating SDOH Was Used to Inform Patient Care

Characteristics	Unadjusted OR^{^a} (95% CI)	Adjusted OR^{^b} (95% CI)
Clinicians
Credentials
Physician	ref	…
Advanced practice provider^{^c}	2.18 (0.71-6.66)	…
Sex
Female	ref	…
Male	0.56 (0.21-1.45)	…
Encounters
Type
Office visit	ref	…
Telehealth	0.82 (0.27-2.54)	…
Reason for visit
Acute care	ref	ref
Behavioral health	0.45 (0.14-1.55)	0.84 (0.23-3.14)
Chronic condition	0.77 (0.32-1.85)	1.00 (0.40-2.44)
Hospital follow-up	3.04 (1.18-7.80)	4.82 (1.62-14.67)
Routine care	1.03 (0.59-1.82)	1.35 (0.73-2.46)
Insurance type
Medicare	ref	ref
Medicaid	1.65 (0.83-3.26)	1.24 (0.69-2.22)
Private	1.17 (0.65-2.11)	1.15 (0.63-2.12)
Uninsured	3.79 (1.67-8.61)	1.84 (0.84-4.02)
SDOH in EHR at encounter?
Yes	2.62 (1.45-4.72)^{^d}	3.18 (1.79-5.66)^{^d}
No	ref	ref
Pre/post–COVID-19 pandemic^{^e}
Pre	ref	ref
Post	0.49 (0.18-1.31)	1.03 (0.35-3.07)
Patients
Sex
Female	ref	ref
Male	1.50 (0.98-2.29)	1.75 (1.18-2.59)^{^d}
Age, y
<18	ref	–
18-64	0.62 (0.20-1.97)	–
≥65	0.42 (0.13-1.42)	–
Race/ethnicity
Non-Hispanic White	ref	ref
Other^{^f}	2.69 (1.34-5.40)^{^d}	1.50 (0.92-2.45)
Preferred language
English	ref	ref
Other	3.64 (1.45-9.14)^{^d}	3.49 (1.42-8.54)^{^d}

EHR = electronic health record; OR = odds ratio; SDOH = social determinants of health.
↵a All models account for correlation within provider using a robust sandwich error adjustment.
↵b Model adjusted for all covariates that had a univariate P value of ≤0.20 (SDOH screening in EHR, patient sex, race/ethnicity, preferred language, and pre/post COVID-19); model accounts for correlation within clinician using robust sandwich error adjustment.
↵c Nurse practitioner or physician assistant; 1 recruited clinician was a registered nurse.
↵d Significant at P ≤.05.
↵e Pre–COVID-19 pandemic for surveys completed before March 2020, post–COVID-19 for surveys completed after March 2020. No surveys were completed in March 2020.
↵f Includes Hispanic/Latine ethnicity, Black, Asian, American Indian/Alaska Native, multiple races, and unknown race.