Professional Documents
Culture Documents
Contraception
journal homepage: www.elsevier.com/locate/contraception
a r t i c l e i n f o a b s t r a c t
Article history: Objective: Person-centeredness is a critical component of quality in family planning. We previously val-
Received 20 July 2020 idated an 11-item Interpersonal Quality of Family Planning (IQFP) scale. We sought to create a parsimo-
Received in revised form 13 January 2021
nious version of the scale in preparation for testing its appropriateness as a patient-reported outcome
Accepted 20 January 2021
performance measure.
Study design: To explore clarity and importance of each of the 11 items, we conducted English and
Keywords: Spanish cognitive interviews with patients who received contraceptive counseling (n = 33) at 3 publicly
Contraception funded California clinics. We triangulated these results with psychometric analysis of previously collected
Contraceptive counseling IQFP data (n = 1097) to assess validity and reliability of selected item combinations.
Patient-centered
Results: The 11-item IQFP scale was reduced to a 4-item scale (the Person-Centered Contraceptive Coun-
Patient-reported outcome measure (PROM)
seling scale, or PCCC) that includes items evaluating provider performance regarding respect for patients,
Patient-reported outcome performance
measure (PRO-PM) information provision, and eliciting and honoring patient preferences for birth control. Interview partic-
Person-centered ipants deemed the items included in the 4-item PCCC important and clear in both English and Spanish
versions of the instrument. The 4-item PCCC retained the 11-item IQFP’s psychometric properties, in-
cluding internal consistency (Cronbach’s alpha = 0.92 vs 0.97 for the PCCC and IQFP, respectively) and
a consistent single factor analysis solution (factor loadings = 0.86–0.92 and 0.81–0.91). The 4-item PCCC
additionally retained the construct and predictive validity of the IQFP.
Conclusions: The 4-item PCCC is a valid and reliable as a measure of person-centered contraceptive coun-
seling that reflects patients’ perspectives on contraceptive counseling.
Implications: Person-centered measures such as the 4-item PCCC can help inform efforts to improve
health care quality. Future work will investigate the validity and reliability of the 4-item PCCC as a per-
formance measure to determine the appropriateness of its use in the quality improvement context.
© 2021 Elsevier Inc. All rights reserved.
1. Introduction riences, preferences for care, and preferred outcomes can therefore
be challenging to determine.
The measurement of health service quality is a growing pri- Designing quality measures focused on patient-centeredness—
ority across health care, including in the field of family planning. which by definition takes into account patient preferences—is im-
Defining quality in contraceptive care is a uniquely complex issue, portant in all aspects of health care, with patient-centeredness
as patients’ decisions around pregnancy, childbearing, and use of recognized by the National Academies of Medicine as a core do-
contraceptive methods are highly individualized and personal [1]. main of health care quality [2]. Given the personal nature of fam-
How best to measure quality across a broad range of patient expe- ily planning care, these types of measures are of particular im-
portance in this context, including the need for measures that
∗
Corresponding author.
E-mail address: christine.dehlendorf@ucsf.edu (C. Dehlendorf).
https://doi.org/10.1016/j.contraception.2021.01.008
0010-7824/© 2021 Elsevier Inc. All rights reserved.
C. Dehlendorf et al. Contraception 103 (2021) 310–315
311
C. Dehlendorf et al. Contraception 103 (2021) 310–315
312
C. Dehlendorf et al. Contraception 103 (2021) 310–315
Table 2
Participant rankings of the Interpersonal Quality of Family Planning (IQFP) scale item importance.
to refer to someone’s demographic characteristics, such as age the combinations examined, predictive validity of the measure was
or race/ethnicity, and expressed disapproval of these characteris- maintained by 5 of the item combinations. Given qualitative and
tics influencing a provider’s advice on birth control. One Spanish- investigator concerns around clarity of items 4 and 6, combina-
speaking participant understood “personal” in this context to mean tions including these items were removed from further consider-
the privacy of various methods. An English-speaking participant ation. The 2 resulting options had nearly identical psychometric
expressed that “personal” was vague in this context, stating: “I properties and thus a final solution was selected on the basis of
don’t really know what my personal situation is … are you talk- parsimony of items.
ing about me being a grad student, a dancer, a person who is dat-
ing? When I read the question, I can’t really think of anything I’d 3.3. Final measure
put under this.” Item 7 was also frequently cited as confusing. Sev-
eral participants said they did not think having a “plan” applied to The PCCC (Table 4) includes 4 items (items 1, 3, 5, and 8) that
them because they would be using a LARC method requiring little were indicated in both qualitative and quantitative analysis to re-
patient effort after insertion. Participants generally understood the tain the IQFP’s measurement of 3 conceptual domains of person-
introductory instructions to the measure. However, the research centeredness in contraceptive counseling (Interpersonal Connec-
team noted that reading level for this text was higher than desired, tion, Information Sharing, and Decision Support) (Table 4). In cog-
and revised this text for the resulting shortened measure. nitive interviews, participants found these items to be evocative of
When asked which of the 11 items might be repetitive, partic- both their overall patient experience and specific exchanges and
ipants often replied that Items 8 and 10, 5 and 6, and 4 and 11 aspects of their appointment that influenced that experience, with
were repetitive item pairs, and that there was overlap between 3 included items scoring among those most important (all items re-
and 4. In making final item combination selections for quantitative ceived a score of < 5.50). All included items were deemed under-
analysis, we avoided including item pairings described as repetitive standable and equivalent between Spanish and English. This com-
by interview participants, as well as those items most frequently bination retained all of the psychometric properties of the IQFP,
described as unclear. Participant rankings of item important are including strong internal consistency (Cronbach’s alpha = 0.92 vs
depicted in Table 2, and highly ranked items were prioritized in 0.97 for the PCCC and IQFP, respectively), a consistent single fac-
resulting item combinations. tor analysis solution (factor loadings 0.86–0.92 and 0.81–0.91 for
Interviews with bilingual participants indicated general equiv- the PCCC and IQFP, respectively), and high item-total correlations
alence of items across language. As emerging results indicated (0.73–0.82 and 0.75–0.84 for the PCCC and IQFP, respectively). Di-
which items would be appropriate for inclusion in the PCCC, we rectly comparing PCCC scored responses (dichotomous score of 20
explored improving the translations on these specific items. Several vs <20) to the 11-item standard score (dichotomous score of 55
bilingual and Spanish-monolingual participants provided recom- vs <55), yielded a specificity of 1.00 and sensitivity of 0.80 for
mendations to improve the clarity of item 5, and bilingual partic- the PCCC. Validity of the PCCC was comparable to the IQFP, with
ipants confirmed these changes improved the equivalence of item all previous significant associations between scored responses with
meaning in English and Spanish. external validity items retained and similar in strength (Table 3).
Both the IQFP and PCCC are associated with measures of con-
3.2. Psychometric analysis struct validity (high global satisfaction with visit, satisfaction with
method selection process, p < 0.001 for IQFP and PCCC, satisfac-
Full descriptions of the 2 samples used in analyses are de- tion with method choice, recommendation of provider to friend, p
scribed elsewhere [22,23]. Two-parameter logistic IRT model re- < 0.01), and predictive validity, with high IQFP and PCCC scores
sults indicated that item responses did not vary substantially predictive of contraceptive continuation at 6 months (p < 0.05).
with regard to difficulty (range 7.64–12.53) or discrimination lev-
els (range −0.72–−0.92), and therefore no items were excluded or 4. Discussion
specifically selected based on the IRT results. Twelve combinations
of 3 to 6 of items were selected for initial psychometric testing on Our iterative mixed-methods approach to the reduction of
the basis of items receiving adequate patient importance ratings the 11-item IQFP scale produced a brief measure of person-
(items 7, 9, and 11 not considered given ratings >6), and adequate centeredness in contraceptive care that retains the validity of the
representation across the three domains. Multiple psychometric in- original scale and reflects patient preferences for counseling. Com-
dicators remained consistently strong across the item groupings in- pared with the IQFP, which was developed for the context of re-
cluding: consistent item-total correlations (all > 0.70), Cronbach’s search, the 4-item PCCC scale is more feasible for use in quality
alpha (0.88–0.94), and factor loadings (all > 0.70). While associ- improvement, where it is critical to minimize the burden of data
ations with construct validity items were overall consistent across collection on patients and administrators. Like the IQFP, the PCCC
313
C. Dehlendorf et al. Contraception 103 (2021) 310–315
Table 3
Construct and predictive validity of the 4-item Person-Centered Contraceptive Counseling (PCCC) scale compared to the 11-item Interpersonal Quality of Family Planning
(IQFP) scale.
Construct validity % High global visit satisfaction 96.2% 61.8% 95.3% 55.0% 14.80 (9.29, 23.60)c 16.33 (10.71, 24.90)c
Construct validity % Completely satisfied with method selection process 83.8% 45.1% 80.9% 41.1% 8.35 (5.25, 13.27)c 7.73 (4.77, 12.54)c
Construct validity % Very/completely satisfied with method choice 69.9% 48.5% 68.2% 46.5% 1.97 (1.14, 3.38)a 2.02 (1.28, 3.19)b
Construct validity % Would recommend to friend 96.0% 63.6% 94.4% 55.7% 14.90 (5.95, 37.32)c 15.56 (6.10, 39.68)c
Predictive validity Contraception method continuation at 6 months 45.6% 36.1% 45.3% 35.0% 1.74 (1.17, 2.59)a 1.58 (1.01, 2.53)a
Adjusted logistic regression odds ratios accounting for clustering by provider (random effect) and clinic (fixed effect), age (continuous), race/ethnicity, federal poverty level.
a
p < 0.05.
b
p < 0.01.
c
p < 0.001.
Table 4
Final Person-Centered Contraceptive Counseling (PCCC) scale.
Think about your visit. How do you think [provider name] did? Please rate them on each of
the following by circling a number. Poor Fair Good Very good Excellent
Respecting me as a person 1 2 3 4 5
Letting me say what mattered to me about my birth control method 1 2 3 4 5
Taking my preferences about my birth control seriously 1 2 3 4 5
Giving me enough information to make the best decision about my birth control method 1 2 3 4 5
For printable versions of the survey in English and Spanish, see the appendices online or visit https://pcccmeasure.ucsf.edu/.
scale can be interpreted dichotomously, with respondents giving resentative by race/ethnicity [15,23], we ensured that item selec-
either a top score of 5 on all 4 items, or less than this top-box tion reflected a large number of patient responses, while still col-
score. A simple metric for person-centeredness can be calculated lecting rich qualitative data to guide the selection process. There
as the percentage of respondents who gave the top-box score on was also potential for social desirability bias in interviews if par-
the PCCC scale at a given level of aggregation (e.g., provider or ticipants perceived that the research team was affiliated with the
facility). As such, the four items of the PCCC scale constitute a providers from whom they received care. While social desirabil-
PROM for person-centeredness and, when aggregated, a PRO-PM. ity may have affected participant responses to the items, it would
As frameworks and recommendations for contraceptive care qual- have less impact on participants’ item interpretations. Our quanti-
ity developed both in the United States and globally define person- tative analyses used the same data source as the validation study
centeredness as a core dimension of contraceptive care quality, this for the original IQFP, which serves as a strength in our ability to
metric can provide actionable data to improve contraceptive care compare shorter and longer versions of the measure directly, but
services [5–7]. also a limitation in that our analysis here did not demonstrate re-
The PCCC is designed to be used to assess the quality of care producibility.
in any organization that offers contraceptive counseling to its pa- The PCCC provides a valid and reliable approach to measur-
tients in either a family-planning-specific or primary care setting. ing person-centeredness in contraceptive counseling that can be
These settings may include public health department clinics, inde- implemented in a range of care settings, including primary care
pendent nonprofit care providers, Community Health Centers, pub- and specialized reproductive health clinics. Development of PROMs
lic or private hospitals, or other private entities. In addition to its such as the PCCC provides an opportunity to ensure that patient
potential for use in research as a parsimonious measure of patient voices and preferences are prioritized in clinical and translational
experience, if found to be valid and reliable as a PRO-PM, the PCCC research and in quality improvement.
can offer these organizations a unique opportunity to measure the
quality of their patients’ experience of contraceptive counseling.
The use of this measure in quality improvement efforts can include Declaration of competing interest
both as a stand-alone measure of person-centeredness and along-
side contraceptive provision measures as means to track changes The authors declare that they have no known competing finan-
in person-centeredness with changes in provision. This may be of cial interests or personal relationships that could have appeared to
particular benefit in Community Health Centers, where there has influence the work reported in this paper.
been increased focus on building capacity for family planning care
in the years following passage of the Affordable Care Act [25–
29]. Additionally, as many health care systems move toward value- Funding
based payment systems, the PRO-PM could be tied to payment
for improvements to the quality of contraceptive counseling. The The research reported in this publication was funded through
implications for increased attention and value placed on person- the Office of Population Affairs (grant number 1 FPRPA006062-01-
centeredness in contraceptive counseling has the potential to coun- 00) and an anonymous foundation.
teract, or at least call attention to, the ongoing legacy of implicit
racist bias and coercion in contraceptive care [4].
Acknowledgments
Limitations of this work include that our cognitive interviews
informing item reduction were conducted with a relatively small
This work would not have been possible without the input of
number of participants, and only 9% of these participants were
our Patient Stakeholder group, which provided feedback on inter-
Black, limiting generalizability. By triangulating interview data
view guides, IQFP items, survey design, and implementation work-
with a larger sample of quantitative IQFP data that was more rep-
flows.
314
C. Dehlendorf et al. Contraception 103 (2021) 310–315
315