Patterns and Determinants of Patient-Delivered Therapy Uptake Among Healthcare Consumers

Abstract

Background: 

Patient-delivered partner therapy (PDPT) may be an effective alternative to traditional approaches to sexually transmitted infection control and prevention. Consumer PDPT uptake is comprised of two behaviors: (1) patient willingness to deliver medicine (patient-delivery) and (2) partner willingness to take it (partner-use).

Methods: 

505 participants aged 18 to 47 completed a nationally distributed survey of intentions, attitudes, perceived norms, perceived control, and specific outcome beliefs, normative influences, and barriers/facilitators related to PDPT. Zero-order correlations and multivariate statistics were used to describe PDPT uptake patterns and determinants.

Results: 

Consumer uptake exhibited a highly polarized response pattern with the majority willing to participate in patient-delivery (83%) or partner-use (69.4%). Psychosocial determinants of uptake varied by behavior. Regression models of the psychosocial variables (Adj. R²>.75) indicate that, across behaviors, PDPT uptake was most closely associated with perceived norm, followed by attitude and perceived control. Specific factors related to uptake include: beliefs that PDPT makes curing disease easier, reduces the spread of the disease, and ensures people are cured; recommendation by healthcare providers; and that the medicine comes with a note from the healthcare provider and is sealed. Several participant characteristics were also significantly correlated with uptake.

Conclusions: 

Consumers are willing to engage in PDPT and perceive delivery and use differently. Several beliefs and participant characteristics emerged as potentially important factors in willingness to participate. This information can inform the passage of enabling legislation; the selection of PDPT participants; and the development of provider, patient, and partner informational materials.

In Brief

A nationally distributed survey found strong consumer willingness to engage in patient delivery and partner use of patient-delivered partner therapy and identified personal, social, and environmental determinants of such uptake.

THE COMMON APPROACH TO TREATING individuals diagnosed with a bacterial sexually transmitted infection (STI) is to treat the presenting patient—the “index patient”—and to ask the patient to notify his or her sexual partners of the infection so the partners can seek treatment. In some situations, an attempt is made by the healthcare system itself to notify the patient’s partners directly through various forms of provider referral, including partner tracing efforts by disease intervention specialists. This referral system works (although imperfectly) for STI such as syphilis; it is less feasible for more common STI such as Chlamydia.¹

Patient-delivered partner therapy (PDPT) and its umbrella term, expedited partner therapy (EPT), are components of a broader approach to prevention and treatment of STI termed expedited partner services (EPS). In this approach, services usually provided directly by a health care professional are instead “outsourced” to patients themselves. Under the PDPT strategy, when a healthcare provider determines that an index patient is a candidate for participation in PDPT and the patient agrees to participate, the patient is given either a prescription or additional prescription medicine to deliver to his or her partners. The partners are not required to visit the healthcare provider in order to receive treatment, thereby “expediting” the services. STI PDPT represents a natural extension of both partner notification and presumptive treatment and seeks to increase treatment among sexual partners, thereby reducing incidence and prevalence of infection.

Some EPT, for example the prophylactic treatment of meningococcal exposure, is fairly established and acknowledged.² The practice is less established, and more legally ambiguous,^3,4 for treating the sexual partners of individuals with nonviral STI diagnoses;^2,5,6 nevertheless, it occurs fairly frequently.^5,6 Moreover, there is movement to further establish STI PDPT as a legitimate health service - including recommendation by the CDC that PDPT be implemented in certain clinical situations for uncomplicated gonorrhea and Chlamydia infections.²

Several randomized controlled trials have supported patient willingness to deliver medicine to partners and have reported partner compliance with medication instructions, leading to reduced recurrence, particularly among patients with Chlamydia or gonorrhoea.^2,7–11 Because clinical and consumer adoption of PDPT rely on engagement in health-related behaviors which are likely to be associated with social stigma and to be emotionally charged at an interpersonal level, it is essential to have an understanding of the factors that influence PDPT uptake. Although focus groups and other qualitative research efforts limited to clinical populations have identified several factors, including perceptions of the medication packaging, partner trust, and the nature and quality of the relationship,^7,12 little empirical research has directly examined consumer uptake of PDPT medication delivery and use behaviors or the personal, social, and environmental determinants of uptake.²

To address this gap, the current research drew upon a reasoned action approach to understanding health behavior.^13–15 This sociocognitive approach posits that people’s health-related behaviors can be predicted based upon knowledge of their willingness and intentions to engage in the behavior. Willingness itself can be predicted reasonably well if a person’s beliefs about the consequences of performing the action (outcome beliefs), about what other people think about the action (normative beliefs), and about things that make it easier or harder to do the action (facilitators/barriers) are understood.¹⁶

The research effort reported here focuses on two components of PDPT uptake:

1. Willingness of patients to deliver medicine to a partner (patient-delivery/deliverers).
2. Willingness of partners to take a medicine delivered to them (partner-use/receivers).

The research sought to answer the following questions:

1. Are people willing to engage in PDPT?
2. Does PDPT uptake vary by component behavior (patient-delivery vs. partner-use)?
3. What psychosocial constructs are most associated with PDPT uptake?
4. What specific consumer beliefs are most associated with PDPT uptake?
5. What other individual beliefs might influence PDPT uptake?
6. What specific participant characteristics are correlated with PDPT uptake?

Methods

Researchers approached individuals in public locations and invited them to complete a questionnaire regarding sexually transmitted infections and their treatment.

Instruments

Equivalent survey instruments, with wording changes to specify the nature and context of the behavior (delivering or receiving), were developed for the patient-delivery and partner-use behaviors. Overall PDPT uptake (willingness), attitudes, perceived norm, and perceived control regarding patient-delivery and partner-use were assessed with three items each. Outcome beliefs were assessed with an outcome evaluation item (i.e., how good or bad is the outcome) and an outcome likelihood item (i.e., how likely is the outcome). Normative influences were assessed with an imputed belief item (i.e., would a particular normative influence believe I should do something) and a motivation to comply item (i.e., how likely am I to want to do what the normative influence wants). Facilitators/barriers were each assessed with a single impact item (i.e., to what extent would the facilitator/barrier make performance more or less likely). All psychosocial and belief items were assessed on seven point agreement, likelihood, or semantic differential items. Participant characteristic data were also gathered (Table 1). English and Spanish versions of the instruments were field tested among general population- and low-literacy participants before study implementation.

TABLE 1: Participant Characteristics

Participants

505 individuals aged 18 to 47 (M = 30; SD = 8.5), geographically distributed among 11 localities in the United States, participated in the study (patient-delivery version, 260; partner-use version, 247) in August 2007. Persons who did not clearly appear to be children were approached in public places (e.g., parks, malls, and outside public aid and public health offices), screened on age and employment in market research, and invited to participate. The sampling technique was time-space. No active sampling strategies were used (such as oversampling or targeting) to determine who was approached; however, geographic locations were selected based on the likelihood of diverse participation. No significant participant characteristic differences were identified between the patient-delivery and partner-use samples.

Procedure

Each participant was approached, asked to complete a questionnaire regarding STIs and their treatment, and provided an informed consent document describing the research. The refusal rate was approximately 20%. Because of length, completing both instruments would have placed undue burden upon the participant, therefore, individuals who agreed to participate were given either the patient-delivery or partner-use instrument. The respondent completed the questionnaire independently, returned it to the research personnel, and was provided a nominal stipend. Twelve percent of the questionnaires were completed in Spanish.

Scale Construction and Data Coding

All rating items were recoded from the original 1 to 7 range to a range of −3 to 3 to convey the positive and negative nature of the ratings. Uptake, Attitude, Perceived Norm, and Perceived Control scales were calculated as the mean of the three items representing each psychosocial construct (Cronbach’s Alpha >0.78 for each scale). A single weighted belief item was created for the outcome beliefs and normative influences by multiplying the two component items. The facilitator/barrier items were used as gathered. Calculated items and scales were transformed to maintain the −3 to 3 range.

Analysis

Five stages of analysis, similar to those suggested by Fishbein and colleagues for the identification of psychosocial determinants of particular behaviors,^17,18 were conducted to describe the outcomes, identify models, and examine the influence of participant characteristics. The analyses included descriptive statistics, multiple regression and correlation (MRC) of uptake on psychosocial constructs, MRC of uptake on individual beliefs, and additional zero-order correlational analyses of uptake with individual beliefs and with participant characteristics. These analyses were conducted for both (1) a patient-delivery model, capturing aspects of deliverer uptake of PDPT (deliverers); and (2) a patient-use model, capturing aspects of receiver uptake of PDPT (receivers). SPSS (Version, 15, SPSS Inc, Chicago, IL) was used for all analyses.

Results

Response dispersion for deliverer and receiver uptake was clustered at the endpoints, indicating that most participants in the study either strongly would or strongly would not engage in PDPT, with very few individuals responding in the middle of the range (Fig. 1). A large majority of participants expressed willingness to engage in patient-delivery: 83% expressed some level of willingness (>0 on −3–3 scale) and 45.8% expressed strong willingness (3 on the −3–3 scale). Willingness to engage in partner-use was also high, although less so than willingness to deliver, with 69.4% expressing some willingness to receive/use and 27.9% expressing strong willingness. This difference between delivery and use uptake was significant (t(505) = 4.627, P = 0.000) indicating that participants are more willing to engage in PDPT delivery than they are to engage in PDPT use. Attitude and perceived norm were also statistically significantly higher among deliverers than receivers (Table 2).

Fig. 1: Frequency of participant willingness to engage in PDPT as a receiver or deliverer.

TABLE 2: Psychosocial Constructs, by Group: Means and Standard Deviations

The patient-delivery and partner-use regression models were both significant, accounting for approximately 78% of the variance in uptake (Fig. 2). Attitude, perceived norm, and perceived control each emerged as significant contributors in both models. Perceived norm is the psychosocial construct most predictive of uptake (standardized β ≤0.69), while attitude (β ≤.21) and perceived control (β ≤.11), have less weight across models. Although these uptake-psychosocial construct models do not appear different, belief-level MRC identified significant differences between groups (Fig. 3).

Fig. 2: Regression models of psychosocial constructs, by group.

Fig. 3: Regression models of PDPT Patient-Delivery and Partner-Use Uptake on individual beliefs.

For patient-delivery, the uptake-outcome belief model for patient-delivery accounted for 36% of the variance and yielded two significant predictors: the belief that PDPT would make it easier to cure a partner’s STI (β = .27, P = 0.002) and that it would ensure that the partner was cured (β = .21, P = 0.016). The patient-delivery PDPT uptake-normative influence model accounted for 14% of the variance and yielded two significant predictors: the health care provider (β = .27, P = 0.000) and casual partners (β = -.29, P = 0.000). Finally, the uptake-facilitators/barriers model accounted for 46% of the variance and identified three significant predictors of engagement: if the prescription medicine is sealed, if the prescription medicine is accompanied by a note from a health care provider, and if the medicine comes in an “official” package (β = .29, P = 0.002; β = .22, P = 0.001; and β = .21, P = 0.022; respectively).

The partner-use uptake-outcome belief model identified three significant predictors accounting for 47% of the variance: beliefs related to the risk and severity of allergic reactions (β = .20, P = 0.004), ensuring that the partner, him- or herself, was cured (β = .19, P = 0.018), and reducing the spread of disease to others (β = .17, P = 0.016). The partner-use uptake-normative influence model accounted for 22% of the variance, with the health care provider (β = .44, P = 0.000) as the sole significant predictor. The uptake-facilitators/barriers model accounted for 33% of the variance among receivers with the presence of instructions being the only significant facilitator/barrier predictor of uptake among receivers (β = .27, P = 0.011).

Although MRC points to the most parsimonious predictors of an outcome variable, zero-order correlation data are also useful for understanding health behavior such as PDPT, especially when items may be highly correlated or when the goals include fostering behavior adoption or behavioral change.^17,18Table 3 presents the means, standard deviations, and correlation coefficients (with uptake) for each of the individual outcome beliefs, normative influences, and facilitators/barriers. Deliverer beliefs that engaging in PDPT would make it easier for the partner to be cured, protect one’s own health, protect one’s partner’s health, and ensure that one’s partner is cured, are moderately correlated (r >.4) with willingness to deliver the medicine to a partner. The results for partner-use behavior are similar, although not identical.

TABLE 3: Individual Beliefs, by Group and Overall: Means, SD, and Correlations with Uptake

TABLE 3: (Continued)

For normative influences, the belief that the healthcare provider would want participants to engage in PDPT had the highest mean and correlation with PDPT uptake both within and across groups. The primary partner normative influence had a mean indicative that participants thought that primary partners would want them to engage in PDPT; however, this was correlated with uptake only weakly for the receivers and not at all for the deliverers. The casual partner normative influence was inversely related to engagement for deliverers and uncorrelated for receivers.

The facilitators most highly correlated (r >.4) with uptake, for both patient-delivery and partner-use, were: having the medicine sealed, having it in an official package, having instructions provided, having a note from the health care provider, and the presence or absence of partner trust. The partner being a one night stand was significantly and inversely correlated with PDPT patient-delivery uptake; no barriers were identified for receivers.

Participant characteristics positively correlated with patient-delivery PDPT uptake were: current relationship satisfaction, self-assessed health literacy, income, and having previously borrowed prescription medicine from friends. Characteristics that were inversely correlated with patient-delivery were: Spanish as the survey language, Spanish spoken in the home, and the participant’s self-reported number of current partners. Among receivers, the only characteristic significantly associated with uptake was being of Asian descent; this association was negative. No other participant characteristic, including age, was significantly related to either patient-delivery or partner-use behaviors.

Discussion

Overall, the majority of participants expressed strong willingness to participate in PDPT delivery or use. Participants also had positive attitudes, perceived norm, and perceived control regarding PDPT. These findings suggest that most people would engage in PDPT if the option were offered. Such apparently strong public support of PDPT should inform policy- and decision-makers in the establishment of enabling laws and regulations. Similarly, practitioners should realize that, in general, patients would be highly receptive to PDPT.

The results also suggest the importance of recognizing the two component PDPT behaviors—delivering prescription medication to a partner and receiving prescription medication from a partner—are perceived differently among participants. Participants are more receptive to delivering medicine to a partner than they are to receiving medicine from their partner; they also hold more positive attitudes and have a more positive perceived norm regarding such delivery of medicine as compared to use of delivered medicine. Accordingly, supporting patients in getting partners to accept and use medication may be a more challenging issue than fostering initial adoption among the index patients themselves. Moreover, this difference between deliverers and receivers could lead to prescription medication being delivered and going unused. Perhaps patient and/or partner instructions should include information about what to do with the medicine in the event it is not used.

Differences between patient-delivery and partner-use were also identified in the individual outcome beliefs, normative influences, and perceived facilitators/barriers associated with the psychosocial constructs. Two factors merit additional discussion because the results may be unexpected. First, people’s beliefs regarding allergic reactions occurring during PDPT are positively correlated with partner-use PDPT uptake. This seems to imply that people who think allergic reactions are good also want to participate in PDPT; however, it seems unlikely that people would view an allergic reaction as a good thing. In fact, very few respondents indicated that an allergic reaction was a positive outcome. As with all of the outcome beliefs and normative influences, the allergic reaction belief is a weighted belief: it is comprised of both a likelihood component and an evaluative component.¹⁹ If people believe that allergic reactions are bad but that they are unlikely (an overall positive allergic reaction belief), then they are more likely to take the medicine. Conversely, if people believe that allergic reactions are bad and that they are likely to occur (an overall negative allergic reaction belief), then they are less likely to take the medicine. Consequently, the observed result suggests that reinforcing the unlikelihood of an allergic reaction as part of PDPT medication use may increase partner-use implementation.

A second result of particular interest is the inverse correlation for the normative influence of casual partners for patient-deliverers. This inverse correlation arises primarily from the motivation to comply component of the weighted belief: participants were less likely to engage in PDPT if they were less motivated to do what a casual partner wants them to do, irrespective of whether they believed the casual partner would want them to engage in PDPT. This result was only evident for patient-delivery. Partner-use may not be associated with the casual partner norm because accepting medicine is a fairly passive act. The decision to deliver medicine, on the other hand, entails actively approaching a casual partner, which is likely to be more socially and emotionally difficult. The finding that the partner being a one night stand is a barrier to patient-delivery but not to partner-use supports this interpretation.

Although the results suggest that engagement in PDPT is not a monolithic behavior and should be considered in its constituent parts, a number of factors emerged as similar across the two behaviors. The uptake-individual belief models indicate that capitalizing on or reinforcing beliefs that PDPT makes curing disease easier, reduces the spread of the disease, reduces perceptions of allergic reaction likelihood, and ensures that people are cured, may improve PDPT uptake. In addition, perceived endorsement by healthcare providers and, if a casual partner is involved, higher levels of motivation to comply with casual partner wishes will likely increase adoption. Finally, when the PDPT prescription medicine is sealed, in official packaging, comes with a note from the health care provider, and has accompanying instructions, adoption is likely to increase.

The response rate of approximately 80% is commensurate with similar studies conducted by the research group.^20,21 However, no data were gathered regarding those who refused participation; therefore those who refused may be qualitatively different from those who did not. Although efforts were made to ensure representativeness of the sample, the overall number of participants remains relatively small and the number of tests conducted large, therefore the study should be considered exploratory.

The sample is a nonclinical one. Because PDPT currently occurs mostly in clinic settings, many individuals in the sample may never encounter the situation. However, if PDPT implementation increases, then increasing numbers of the general population will be asked to deliver medicine or have partners bring them medicine. Moreover, comparing the uptake and beliefs of those who had previous clinic experience to those without such experience did not identify any relevant significant differences, indicating that the clinical and nonclinical samples are likely similar in their uptake, attitudes, and beliefs.

The findings revealed some potentially important differences between those who speak English as a primary language and those who speak Spanish; it is possible that these differences are due to language translation issues rather than to underlying differences in uptake or beliefs. A careful translation process, including back translation, partially guarded against this limitation; however, it remains possible that differences in meanings assigned to survey words, for example on semantic differential items, may cloud the results. Although no significant differences were found between those who stated that they spoke Spanish in their homes yet completed the interviews in English and those who stated they spoke Spanish in their homes and completed the interviews in Spanish, providing support for instrument equivalence, the small numbers of individuals in these two groups suggest a Type 2 error could be committed. Regardless, the possibility that those who speak Spanish as a primary language are less likely to take part in PDPT than those who speak English is of potential importance, and no work to date has examined whether PDPT consumer behaviors vary by ethnicity, indicating additional research in this area may be worthwhile.

Finally, the study examines proxy measures of the patient-delivery and partner-use behaviors through willingness and intentions rather than direct measures of behavioral outcomes. People’s expressed willingness to do something in response to a hypothetical question does not necessarily map directly to their actual behavior; however, such intentions have been shown in many studies to accurately reflect future performance of a behavior.¹⁵

Conclusion

People appear strongly willing to participate in PDPT delivery and use, with uptake for delivery appearing to be somewhat greater than uptake for use. As a result of this differential uptake, clinical implementation and future research should consider the component behaviors separately in order to better understand and address the differences. Several beliefs were identified as significant predictors of PDPT uptake. Examination of the use of these beliefs for candidate selection, for behavioral change efforts to increase adoption, or for educational efforts to improve implementation will be a valuable extension of the present work. In addition, uptake appears to vary by several participant characteristics, and future work should more closely examine, for example, whether PDPT uptake varies by spoken language (or ethnicity) or whether delivery or use vary independently by relationship satisfaction. Overall, the majority of the public appears to support PDPT and, moreover, if PDPT were offered, it appears that people would in fact deliver and use medication. Therefore, the issues become supporting passage of enabling legislation in various jurisdictions and further facilitating clinical adoption and implementation.