gms | German Medical Science

Background: As integrated care (IV) becomes important for modern medicine, measuring how IV impacts patient experiences is crucial. The US-developed “Patient Perceptions of Integrated Care” (PPIC) questionnaire was translated and adapted to the German language and health care context (PPIC-Deutsch, PPIC-D).

Question: In this validation study, we assessed acceptability and feasibility, factorial validity, and known-group validity of the PPIC-D with German patients.

Method: Three versions of the PPIC-D – ambulatory care (PPIC-Deutsch-Hausarztversion, PPIC-D-HA), treatment center specific, and condition specific – were field tested with ambulatory patients, patients receiving care at a back-pain clinic, and injury patients enrolled in physical therapy, respectively.

The validity study conveniently sampled from the same patient groups. Ambulatory patients were recruited from primary practices taking part in the IV project Opti-Mum and other ambulatory practices in the same region. Patients were recruited from a back-pain clinic with a comprehensive IV program: they completed questionnaires at the beginning of treatment and again 3-4 weeks later. Lastly, patients undergoing rehabilitation after a physical injury were recruited from several clinics. Some of these patients were treated under employee accident insurance, which we hypothesize has better IV than other insurances.

Distribution and missing data were analyzed to assess feasibility and acceptability. The PPIC-D’s factorial structure was examined using exploratory factor analysis (EFA). We also examined whether the PPIC-D distinguishes across groups known to differ in level of IV: Opti-Mum, patients treated at the back-pain clinic, and injury patients with employee accident insurance (high IV) were compared to other ambulatory patients, back-pain patients before treatment at the clinic, and injury patients with other insurances (low IV). Center and condition specific PPIC-D were combined in analyses (PPIC-Deutsch-Facharztversion, PPIC-D-FA).

Outcomes: Based on field test interviews (ambulatory=20, back-pain=11, injury=14), the PPIC-D’s introductory texts and skip patterns were clarified. No substantive changes were suggested, showing acceptability by German patients. The validation study data collection is completed for ambulatory (n=210) and back pain (n=99) patients, and ongoing for injury patients (n=150).

Distribution: Only 3 PPIC-D-HA items had more than 5% missing and all items fell under 6.5%, demonstrating acceptability and feasibility. Five PPIC-D-FA items had over 5% missing, the highest at 8.5% and 11.8%. Both items demonstrated low levels of missing in the PPIC-D-HA. Percent of positive responses were high for items concerning quality of care, but lower for items regarding specific activities (e.g. doctor asking about private life). This phenomenon was also reported by the original PPIC study by Singer et al. [1].

Factorial Validity: Four factors were extracted from the items that are shared between all three PPIC-D versions. These components roughly correspond to patient-centered care, competency of care, information transfer, and holistic care. These factors do not precisely correspond to the original PPIC’s 7 dimensions.

Known-group Validity: Two items asked patients to rate, from 0 to 10, 1) their entire course of medical care and 2) organization of this care. These items differed significantly across Opti-Mum (means 7.7, 7.3) and other ambulatory patients (means 7.1, 6.7); and back-pain patients pre (means 6.7, 6.0) and post treatment (means 9.1, 9.1). Four other items differed significantly across ambulatory and back-pain patients. Items did not distinguish between injury patients based on their insurance. Based on EFA results, two equally weighted summary scores were assessed. One (patient-centered care) significantly differed across ambulatory and back pain patients while the other (information transfer) differed only across ambulatory patients.

Discussion: The validation study showed that the PPIC-D is accepted by German patients and can distinguish between different levels of IV. Its factorial structure is different from the original PPIC, most likely due to the different systems of health care delivery and the modifications made for the PPIC-D. Two overall care items, four specific items, and one summary score distinguished between patients hypothesized to differ in levels of IV. However, a more sophisticated scoring system must be created. The next step is to test the PPIC-D in broader patient populations.

Practical implications: The PPIC-D can be used to measure patient experiences with IV. The three versions can be adapted for various clinical settings. As we found the instrument, on an item level, to distinguish between patients with different levels of IV, including pre and post treatment, it can be applied cross-sectionally and longitudinally to evaluate IV interventions.