Explanation & Elaboration

13b. Outcomes

Changes in processes of care and patient outcomes associated with the intervention

1. Presents data on changes observed in the care delivery process
2. Presents data on changes observed in measures of patient outcome (for example, morbidity, mortality, function, patient/staff satisfaction, service utilization, cost, care disparities)
3. Considers benefits, harms, unexpected results, problems, failures
4. Presents evidence regarding the strength of association between observed changes/improvements and intervention components/context factors
5. Includes summary of missing data for intervention and outcomes

Example

"Study Flow
.... Of the 205 providers, 23 were subsequently excluded after randomization because patients did not consent (n=222) or because chart review showed that a patient was taking more than 1 medication (n=264). ...At trial completion, 975 patients (73%) had at least 1 follow-up blood pressure reading, including 255 of 324 (78.7%) in the provider education only group, 362 of 547 (66.2%) in the provider education and alert group, and 358 of 470 (76.2%) in the provider education, alert, and patient education group.

Outcome Measures: Systolic and Diastolic Blood Pressure
... the proportion achieving goal blood pressure {SBP <140 mm Hg} differed in the 3 groups: 107 of 255 (42.0%) versus 148 of 362 (40.9%) versus 213 of 358 (59.5%) (P=0.003) in the provider education; provider education and alert; and provider education, alert, and patient education groups, respectively. ... There were no between-group differences in this secondary outcome {DBP<90 mm Hg} (P=0.81) ...

Process Measures: Intensification of Antihypertensive Regimen and Adherence
...[I]ntensification of antihypertensive medications was done in 32.4% of patients in the provider education only group, 28.5% in the provider education and alert group, and 29.1% in the provider education, alert, and patient education group (P=0.81) ... Medication adherence (pharmacy refills) was also measured after intervention, and there were no differences in medication adherence score among study groups ... (P=0.71).

... Death during Follow-up
During the study period...1.1% participants died (8 [2.5%] in the provider education only group, 3 [0.6%] in the provider education and alert group, and 4 [0.9%] in the provider education, alert, and patient education group; P=0.027)." [15]

Elaboration

In addition to providing specific details of the setting, intervention(s), and project evolution, it is equally important to describe changes in patient care and outcomes that occurred in response to the project. Data related to specific process and outcomes measures can be reported as shown in the above example. The study flow diagram (Figure 2) combined with the text provides a clear picture of the study design at a glance and lay out the flow of participants over time. [35]  Providing this level of detail - in combination with a description of the evolution of the changes - helps readers to determine generalizability, or external validity, of the study.

Criteria to determine external validity for implementation studies have been suggested by Glasgow et al. and include the following [36]: (1) representativeness of the sample (target audience, inclusion and exclusion criteria, participation, settings, individuals); (2) program or policy implementation and adaptation (consistency, staff expertise, program customization); (3) outcomes for decision making (significance, adverse consequences, program intensity, costs, moderator effects-such as subgroups of participants/ settings); and (4) maintenance and institutionalization (long-term effects, sustainability, evolution, attrition). Each is demonstrated in the example. In the "Study Flow" section of this example, the drop-out rate and the magnitude of missing data over time are clearly stated. A situational analysis - which was not included in this examples - often includes the actual usage of the intervention, the degree of success in implementing the intervention, and how and why the initial plan for improvement evolved (linked to the initial context description in 4. Local problem, 5. Intended Improvement, and 8. Setting). Such analysis is sometimes considered beyond the scope of the main findings, but can be extremely important in understanding why an improvement intervention did or did not work in a particular setting. Finally, planned interventions that are not fully implemented as intended may result in less reliable findings, so these should also be reported in the manuscript.

Depending upon the study design, results may include means, proportions, standard deviations, risk ratios, confidence intervals, or time-series. Furthermore, unexpected findings (harms or benefits), organizational impact, and difficulties implementing the intervention should be reported as well as the magnitude (effect size) and strength of associations. In the example, the indicated magnitude and strength of association with patient outcomes (Outcome Measures: Systolic and Diastolic Blood Pressure), processes of care (Process Measures: Intensification of Antihypertensive Regimen and Adherence), and other clinical outcomes (Death during Follow-up) are all specified.

References

15. Roumie CL, Elasy TA, Greevy R, et al. Improving blood pressure control through provider education, provider alerts, and patient education: a cluster randomized trial. Annals of Internal Medicine. 2006;145(3):165-175.

35. Egger M, Juni P, Bartlett C, for the CONSORT Group. Value of Flow Diagrams in Reports of Randomized Controlled Trials. JAMA. 2001;285(15):1996-1999.

36. Glasgow RE, Emmons KM. How can we increase translation of research into practice? Types of evidence needed. Annual Review of Public Health. 2007;28:413-433.

 

to top ^