Implementation of disease activity measurement for rheumatoid arthritis patients in an academic rheumatology clinic

The study took place in an academic rheumatology clinic at the VA Hospital in San Francisco. All providers working in the clinic were included in the intervention and analysis. The practice consists of nine rheumatology fellows (two fellows were present for only the pre-intervention period) one nurse practitioner and five attending physicians.

All patients with a diagnosis of rheumatoid arthritis in the medical record (reviewed by AB) were included. They study period lasted from June 2014 until April 2015. This project was considered exempt from IRB approval by the San Francisco VA IRB because it qualified as a quality improvement activity. Providers were aware of the interventions used. No funding sources were utilized for this quality improvement project.

Preliminary work involved interviewing staff, fellows and faculty to understand the barriers to documentation of RA disease activity. Prior to the intervention, a clinician wishing to document a disease activity measure would have to 1) identify RA patients requiring a measure, 2) request this patient’s global assessment on a 0–10 or 0–100 Likert or visual analog scale during the clinical encounter 3) conduct the history and physical, including a 28-joint count, and 4) calculate the CDAI or DAS28 score and remember or look up the corresponding disease activity category 5) document a disease activity measure in the clinic note. No standard method existed for identifying patients ahead of clinic and providing them with a visual analog or Likert scale for rating their patient global assessment. Additionally, a ruler was often not available to score the patient and physician global if a visual analog scale was used. Other obstacles identified prior to the initiation of the first QI cycle were a lack of expectation on the part of the attending physicians to report a disease activity score, a lack of ability to modify the electronic medical record to incorporate the disease activity measures, and time constraints in clinic.

Upon reviewing the barriers to completing disease activity measures, the modifiable obstacles included attending physician expectations and availability of a form on which patients could indicate a global assessment and ruler with which to measure a visual analog scale line. The obstacles that were not immediately modifiable were changing the medical record (CPRS) and the time available to see a patient.

The Clinical Disease Activity Index (CDAI) incorporates a patient global score on a 0 to 10 scale, a physician global score on a 0 to 10 scale and a 28-joint count for both tenderness and swelling. The score is tabulated with up to 28 points for tenderness, 28 points for swelling, 10 points for the physician global and 10 points for the patient global. These scores are then translated into four categories: remission, low disease activity, moderate disease activity or high disease activity [3].

The disease activity score 28 or DAS28 (ESR or CRP) incorporates the patient global score on a scale of 0–100, a tender joint count of 28 joints, a swollen joint count of 28 joints and either the erythrocyte sedimentation rate (ESR) or the C-reactive protein (CRP) and the provider then uses a formula to calculate the disease activity score and it is again characterized as remission, low disease activity, moderate disease activity or high disease activity [3].

A Plan-Do-Study-Act (PDSA) methodology provides a structure of iterative change that adapts to feedback to result in the desired outcome. The first stage, ‘plan’, is a planning stage in which a change is identified with the goal of improvement. The second stage, ‘do’, is when this change is tested. The third stage, ‘study’, examines the success of the change and the ‘act’ stage identifies the next steps needed for beginning the next PDSA cycle [12]. We undertook two PDSA cycles in an attempt to improve use and documentation of disease activity measures at the San Francisco VA rheumatology clinic. There was a 4-week pre-intervention period where we assessed use of disease activity measures prior to any specific intervention. A first PDSA cycle was planned and assessment of its effect on disease activity measurement documentation was planned for 8 weeks later. A second PDSA cycle was planned and assessment of its effect on disease activity measurement documentation was again planned for 8 weeks later. Because of the success of PSDA cycle 2, we continued to follow the outcome for an additional 21 weeks (about 5 months).

The outcome of interest was the proportion of RA patients seen on a given day in clinic who had a documented disease activity measure in the encounter note for that day. Eligible RA patients were identified prior to their clinic visit through medical chart review by one author (AB) based on the assessment sections on the most recent note. Presence or absence of the disease activity measure was determined by review of the entire note.

We divided the study period into three parts – pre-intervention, PDSA cycle 1 and PDSA cycle 2. We created control charts to describe the proportion of encounters on each clinic day with a disease activity measure documented in the clinical note (p-charts). A continuous improvement of 6 or more points in the same direction is considered a trend [13].

During the 4-week pre-intervention period, there were 29 RA patient encounters. Twenty-four percent of pre-intervention visits had a documented disease activity measure in the clinical note. A single clinician was responsible for nearly all of this disease activity score documentation.

The cycle 1 intervention consisted of 1) developing a paper form that would be readily available in clinic for collecting patient and physician global assessments and tabulating joint counts using the homunculus (Additional file 1) and 2) Making the form available in clinic 3) Brief one-on-one educational discussions with incoming fellows on the use of this form, the calculation of CDAI and DAS 28, and the purpose and goal of disease activity measurement. The providers were encouraged to download a CDAI and DAS-28 calculator onto their smartphones [14].

In the workflow (Fig. 1) for cycle 1, providers were to identify RA patients and retrieve a form located in the clinic that included the patient global visual analog scale, physician global visual analog scale, a 28-joint count with homunculus and instructions on how to calculate the CDAI and DAS-28 ESR and CRP. The form included a fold-over ruler, obviating the need for finding a ruler, in addition to instructions for calculating the CDAI.

PDSA cycle 1 lasted 8 weeks. During this period, providers had 59 visits with RA patients, and 26 (44 %) of those encounters had disease activity measures recorded in the note (see Fig. 2).

Follow-up interviews with clinicians identified several issues during cycle 1: 1) the patient and physician global were both completed on one form, so if the patient filled out the global score first, the physician would see the patient global score prior to the recording the physician global score, thus potentially biasing her rating; 2) the global assessment form had to be printed carefully so that the ruler lined up with the patient and physician global lines; 3) since RA patients were not pre-identified prior to their visits, providers would often forget to administer the patient global form.

The cycle 2 intervention involved: 1) the creation of two separate forms so that the patients would complete the patient global in the waiting room, prior to the clinic visit, and the providers would not see the score until after completing their own global assessment; 2) additional education for providers; 3) identification and flagging of RA patients prior to the clinic visit so that patients could fill out global assessments in the waiting room prior to the start of their visit.

The one form used previously was split into two – the first form was designed for patients and included a global assessment scale (see Additional file 2). The second form was for the providers and included a 28-joint count homunculus to record the swollen and tender joints, a physician global score (see Additional file 3). It also included an explanation of how to calculate the CDAI and DAS-28. On both forms, the global scores were changed from a visual analog scale to a numerical Likert scale, noted to be interchangeable in American College of Rheumatology’s working group recommendations [3]. This eliminated the need for a ruler to calculate patient and provider global assessments.

Additional education was directed to providers. This included a post-clinic conference discussion of the history and utilization of disease activity measures and a division-wide Quality Improvement conference in which there was a discussion of the changes at the VA.

RA patients were identified via chart review and flagged by a single author (AB) prior to clinic (see Fig. 3). Appropriate forms were provided to medical assistants prior to clinic. On the day of clinic, the medical assistants were instructed to give the form to flagged patients while taking their vital signs. The patients filled out the form while sitting in the waiting room. The patients would then either hand the form back to the medical assistant, who would attach it to the encounter form given to providers, or the patient could give the form directly to the clinician during the appointment. The physician would then conduct the history and physical, and enter the swollen and tender joint counts and the physician global into the table calculator on Form 2. The physician would then calculate the CDAI and/or DAS28 using simple addition or their IPhone app. When typing the note, the physician was instructed to include the score within the text of the note.

PDSA cycle 2 lasted 27 weeks. During this period, providers had 182 clinic visits with RA patients, and 155 (85 %) of those encounters had disease activity measures recorded in the note (see Fig. 2). Starting during PDSA cycle 2, there were six occurrences over the p-bar, which continued throughout the PDSA cycle, with three exceptions. All providers showed improvement in disease activity measure documentation. There were no outliers.