This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Financial Incentives to Translate ALLHAT Into Practice: A Randomized Trial (NCT00302718), a randomized na study. The primary outcome is Proportion of the Physicians' Patients With Blood Pressure Control or Appropriate Response to Uncontrolled Blood Pressure.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT00302718 |
| Phase | NA |
| Condition | Hypertension |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None (open label) |
| Enrollment | 83 |
| Arms in the primary outcome | n/a |
| Lead sponsor | VA Office of Research and Development |
| Results posted | Yes |
2. Clinical Question
Population
Patients with hypertension treated by participating physicians and practice groups. At baseline, the measure was based on electronic health record review of 40 randomly selected patients with hypertension from each physician’s panel. The registry record says the physician count in this measure refers to those participating during the baseline period; it does not give the final physician count for the repeated-measures analysis.
Intervention
Financial incentives directed to physicians, practice groups, or both. The study examined these approaches separately and in combination as behavioral interventions intended to influence hypertension quality of care.
Comparator
The control group received no incentives. Its physician participants received audit and feedback performance reports, as did participants in the intervention groups. Thus, the comparison concerns incentives in addition to audit and feedback, rather than incentives versus no performance feedback.
Primary question
Whether financial incentives were associated with a higher proportion of physicians’ patients meeting the study’s blood-pressure outcome: achieving recommended blood-pressure thresholds or receiving an appropriate response when blood pressure was uncontrolled. The study hypothesis also referred to prescribing guideline-recommended antihypertensive medications and achieving JNC 7 blood-pressure goals. The registry record does not report the observed outcome values or analysis results, so it does not establish whether the groups differed.
3. Trial Design
The registry describes a randomized, parallel-group trial with no masking. It lists an enrollment of 83 and identifies hypertension as the condition. The study was completed; the registry record gives a start of 2007-02 and primary completion of 2011-10. The sponsor was the VA Office of Research and Development.
The intervention groups received physician-level financial incentives, practice-level financial incentives, or both. The control group received no incentives, but its physicians received audit and feedback performance reports, as did the intervention-arm participants. The comparison therefore concerns whether adding financial incentives to a setting with audit and feedback is associated with differences in care quality.
The record identifies the design as cluster-randomized. In a cluster design, assignment occurs at a group level rather than independently for every patient; patients cared for by the same physician or practice may have more similar outcomes than patients in different clusters. This dependence matters when estimating uncertainty. A simple patient-level analysis that treats every patient as statistically independent could misrepresent precision. The registry’s analysis description says the participant flow diagram gives the final number of physicians who received at least two instances of the intervention and were included in the repeated-measures longitudinal analysis. The actual final count and the details of that analysis are not reported in the registry record.
The registry lists 15 outcome measures and 4 analyses, with secondary outcomes involving CRC screening, LDL cholesterol, Hb A1c, and beta blocker use. The record does not report a detailed randomization procedure, allocation concealment, analysis model, or protocol-based sample-size justification. These omissions limit how fully the design and analysis can be independently assessed from the registry record.
4. Participant Flow and Accounting
The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.
Overall Study
| Stage | Physician-level Incentives | Practice-level Incentives | Physician- and Practice-level Incentives | No Incentives (Control) |
|---|---|---|---|---|
| STARTED | 20 | 21 | 21 | 21 |
| COMPLETED | 19 | 20 | 19 | 19 |
| NOT COMPLETED | 1 | 1 | 2 | 2 |
Reasons for not completing this period:
| Reason | Physician-level Incentives | Practice-level Incentives | Physician- and Practice-level Incentives | No Incentives (Control) |
|---|---|---|---|---|
| Participant left the hospital | 1 | 0 | 0 | 1 |
| Withdrawal by Subject | 0 | 1 | 0 | 0 |
| Participant left primary care setting | 0 | 0 | 2 | 1 |
Analysis population for the primary outcome, as defined in the registry: The number of physicians listed is those who participated in the baseline period. The participant flow diagram describes the final number of physicians who received at least two instances of the intervention and were included in the repeated-measures longitudinal analysis.
5. Baseline Characteristics
The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.
Population: Distribution of Physician Characteristics Across Study Groups
| Characteristic | Unit | Physician-level Incentives | Practice-level Incentives | Physician- and Practice-level Incentives | No Incentives (Control) | Total |
|---|---|---|---|---|---|---|
| Participants analysed | 19 | 20 | 19 | 19 | 77 | |
| Age, Continuous | years | 45 (SD 8) | 48 (SD 8) | 44 (SD 6) | 48 (SD 7) | 46 (SD 8) |
| Sex: Female, Male: Female | Participants | 11 | 8 | 9 | 6 | 34 |
| Sex: Female, Male: Male | Participants | 8 | 12 | 10 | 13 | 43 |
| Race/Ethnicity, Customized: White | participants | 7 | 7 | 9 | 8 | 31 |
| Race/Ethnicity, Customized: Black | participants | 3 | 2 | 1 | 0 | 6 |
| Race/Ethnicity, Customized: Asian Indian | participants | 5 | 5 | 6 | 5 | 21 |
| Race/Ethnicity, Customized: Other | participants | 4 | 6 | 3 | 6 | 19 |
| Region of Enrollment: United States | participants | 19 | 20 | 19 | 19 | 77 |
6. Primary Endpoint
The registry lists the primary outcome as Proportion of the Physicians' Patients With Blood Pressure Control or Appropriate Response to Uncontrolled Blood Pressure, measured over Baseline period (August-November 2007).
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | not applicable |
| Unit | percentage of physicians' patients |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint is the unadjusted proportion of physicians’ patients who met the study outcome during the baseline period, identified as August-November 2007. Meeting the outcome meant that a patient achieved the recommended blood-pressure thresholds or that the provider appropriately responded to uncontrolled blood pressure. The specified responses included increasing the dosage of a guideline-recommended antihypertensive medication or recommending a lifestyle modification for a patient with Stage 1 hypertension.
The endpoint uses electronic health record review. For each physician, records were reviewed for 40 randomly selected patients with hypertension from that physician’s panel. This approach focuses on documented care and patient blood-pressure status, rather than relying only on a clinician’s self-report. It also makes clear that the patient observations are grouped by physician, a feature that matters to statistical analysis.
Blood-pressure control and appropriate response are clinically relevant quality-of-care indicators because they connect measured patient status with documented provider action. Combining control with an appropriate response can capture care for patients whose blood pressure remains uncontrolled, rather than counting only whether a target was reached. However, the registry record gives the endpoint definition and baseline timeframe, not the observed proportion or later-period estimates. It therefore does not permit conclusions about the magnitude, direction, or precision of any incentive-associated change.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry record describes a trial designed to assess financial incentives for hypertension quality of care, but it does not report the primary endpoint values, effect estimates, confidence intervals, or p-values. Consequently, the posted information in this record cannot show whether physician-level incentives, practice-level incentives, or the combined approach were associated with a difference in the primary outcome. The study hypothesis is not a result: it states what the investigators expected to examine, not what the trial found.
The primary measure is an unadjusted baseline-period proportion. A baseline value describes the measure before or at the beginning of the performance evaluation; by itself, it cannot establish an intervention effect. The record explains that the physicians listed for the measure participated in the baseline period, while the final physician count for the repeated-measures longitudinal analysis is described as belonging in a participant flow diagram. That final count is not reported in the registry record. Without the repeated-measures estimates and analysis details, the baseline description cannot be used to infer changes over time or differences between arms.
The control group received audit and feedback reports, as did the intervention groups. Any trial comparison would therefore need to be interpreted as the added association of financial incentives against a background of audit and feedback, not as financial incentives versus absence of all quality-improvement activity. The distinct incentive levels also mean that the trial’s question is not reducible to one comparison. The registry identifies several intervention groups but does not report the estimates for their comparisons with control or with one another.
Cluster randomization is important to interpretation because patients within the same physician or practice cluster may share care processes. Statistical uncertainty should reflect that grouping. The registry record does not provide sufficient model details or estimates to assess how that dependence was handled in the reported analysis. The record states that a simple individual-level re-analysis would be misleading and that only the sponsor’s own analysis is shown. In any case where a recalculated result conflicts with the sponsor’s analysis, the sponsor’s own analysis takes precedence.
The enrollment figure and the numbers listed in the adverse-event groups do not substitute for an outcome analysis. Nor can the absence of reported effect estimates be interpreted as evidence of no effect. The appropriate conclusion from this record is limited: it describes the trial question, design, and endpoint, but the result and its statistical precision are not reported in the registry record.
9. Statistical Methods Explained
Why does the cluster design matter?
In general, cluster randomization assigns groups of people or care settings rather than assigning each patient independently. Patients treated by the same physician can share prescribing habits, workflows, and other aspects of care. Their outcomes may therefore be correlated. When observations are correlated within clusters, treating them as independent can make standard errors too small and confidence intervals too narrow. An appropriate analysis accounts for the grouping when estimating effects and uncertainty. For this trial, the registry explicitly cautions that a simple individual-level re-analysis would be misleading and says only the sponsor’s own analysis is shown. The detailed sponsor model is not reported in the registry record.
What does a repeated-measures analysis address?
In general, repeated-measures analysis considers observations made across more than one time or performance period. It can account for measurements related to the same unit over time, such as a physician’s quality measure across periods. This can distinguish changes over time from differences between groups, subject to the model and available data. The registry describes a repeated-measures longitudinal analysis and says that the final physician count for that analysis is described in a participant flow diagram. The count, period-specific estimates, and model specification are not reported in the registry record. The primary endpoint description itself concerns the baseline period, so it should not be mistaken for a longitudinal treatment-effect estimate.
What does an unadjusted proportion convey?
In general, a proportion is the number meeting a defined criterion relative to the number assessed. An unadjusted proportion is a direct summary that has not been adjusted for other characteristics in a statistical model. It can be understandable, but it does not by itself account for differences among physicians or patients, the cluster structure, or repeated observations. The record defines the measure using patient records sampled from each physician’s panel and labels it unadjusted. It does not report the baseline proportion or an adjusted estimate.
How should uncertainty be read?
In general, a confidence interval describes a range of values compatible with an estimate under the assumptions of the analysis. Its width reflects statistical precision: wider intervals indicate less precision than narrower intervals. A p-value addresses compatibility with a specified null model; it does not measure clinical importance or the probability that a hypothesis is true. Neither confidence intervals nor p-values for the primary endpoint are reported in the registry record. Without estimates and uncertainty intervals, the size and precision of any group differences cannot be assessed.
What do the planning illustrations mean?
In general, sample-size planning calculations illustrate the amount of information that might be needed to detect a specified effect under stated assumptions. The registry’s planning assumptions describe a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect defined as a stated fraction of the observed effect. For time-to-event outcomes, the described calculations concern total events; other calculations concern participants per arm. These are planning illustrations based on an observed effect, not the trial’s own design justification and not observed power. The resulting planning values are not reported in the registry record. No sample-size justification or corresponding protocol calculation is reported in the registry record.
Why does over-dispersion matter for count outcomes?
In general, count outcomes can vary more than a simple statistical model assumes, a property often called over-dispersion. If it is present and ignored, uncertainty estimates may be unreliable. The registry categorizes the primary measure as a count-related outcome type, but its clinical definition is a proportion and no dispersion estimate or count-model specification is reported in the registry record. It would therefore be inappropriate to assert that a particular count model or over-dispersion adjustment was used.
The registry lists 15 outcome measures and 4 analyses, but does not report enough detail to reconstruct the analysis plan, the statistical models, covariate adjustments, or handling of multiple comparisons. The stated design and endpoint support an explanation of the issues an analysis should address; they do not establish which specific procedures were used beyond the descriptions in the registry record.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists 15 outcome measures and 4 analyses. Secondary outcomes include CRC screening, LDL cholesterol levels, Hb A1c levels, and beta blocker use. When a study examines multiple outcomes or comparisons, the chance of at least one apparently notable result can increase if each is assessed separately without accounting for multiplicity. In general, a prespecified hierarchy, adjustment procedure, or clear distinction between primary and exploratory analyses can help readers interpret a collection of results. The registry record does not report a multiplicity adjustment, testing hierarchy, or detailed statistical analysis plan. The listed outcome and analysis counts alone do not reveal which comparisons were prespecified or how results were interpreted.
Missing-data handling is also not reported in the registry record. The primary endpoint description specifies electronic health record review for randomly selected patients, but does not explain how unavailable records, incomplete blood-pressure documentation, or missing follow-up observations were handled. The final number of physicians included in the repeated-measures longitudinal analysis is referenced as appearing in the participant flow diagram, but that number and the flow details are not reported in the registry record. It is therefore not possible to assess attrition or determine whether missingness differed across groups from the information recorded.
The analysis populations are not fully described. The record distinguishes physicians participating at baseline from physicians included in the repeated-measures analysis after receiving at least two instances of the intervention, but does not report a complete definition of the analysis population or how all randomized clusters were handled. It also does not report whether analyses followed assignment regardless of intervention receipt. These details matter because exclusions after randomization can affect comparability. The record’s warning against a simple individual-level re-analysis is particularly relevant: any missing-data or population assessment should respect the cluster structure and the sponsor’s analysis.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Physician-level Incentives | 0 | 19 | 0.0% |
| Practice-level Incentives | 0 | 20 | 0.0% |
| Physician- and Practice-level Incentives | 0 | 19 | 0.0% |
| No Incentives (Control) | 0 | 19 | 0.0% |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Physician-level Incentives | 0 of 19 |
| Practice-level Incentives | 0 of 20 |
| Physician- and Practice-level Incentives | 0 of 19 |
| No Incentives (Control) | 0 of 19 |
The adverse-event record lists four groups: physician-level incentives, practice-level incentives, physician- and practice-level incentives, and no incentives (control). For the groups, the registry lists the numbers at risk for serious and other adverse events and reports 0 affected for each of those event categories. These entries describe the registry’s event counts; they do not provide event definitions, surveillance procedures, follow-up duration, or a detailed safety analysis.
The deaths field is not populated with a numerical value for these groups. Death counts are therefore not reported in the registry record; an unreported value should not be read as a confirmed zero. Similarly, the listed at-risk figures do not establish how completely events were captured or whether all participants were observed for the same period. The registry lists no top serious-event terms.
These safety entries should be read separately from the primary quality-of-care endpoint. They do not quantify the effect of incentives on blood-pressure outcomes and do not establish that an intervention is safe or unsafe. The record gives no basis for comparing event risk beyond the reported counts and listed group figures.
12. What the Result Does, and Does Not, Mean
This trial addresses an important implementation question: whether financial incentives directed to individual physicians, practice groups, or both are associated with hypertension care quality when audit and feedback are also present. Its primary measure combines blood-pressure thresholds with appropriate provider response to uncontrolled blood pressure. That definition recognizes both patient status and documented clinical action.
The registry record does not report the primary outcome results, effect estimates, uncertainty intervals, or the sponsor’s analysis output. It therefore cannot establish the direction, magnitude, or precision of any association between incentive assignment and the endpoint. The investigators’ stated hypothesis should not be mistaken for a finding. The absence of reported results is not evidence that the groups were equivalent or that no effect occurred.
The design also limits what can be inferred from a simple comparison of individual patient records. Randomization was at the cluster level, and the registry cautions that simple individual-level re-analysis would be misleading. In general, cluster-aware analysis is needed to represent dependence among patients treated by the same provider or practice. The registry identifies a repeated-measures longitudinal analysis but does not report its estimates, final physician count, or model details.
The control condition included audit and feedback, so the study question concerns incentives added to that context. The record does not support conclusions about financial incentives compared with no quality feedback, nor does it establish that any one incentive arrangement is superior. Its secondary outcomes broaden the care-quality questions, but results for those outcomes are not reported in the registry record.
The record is informative about the intended comparison, endpoint definition, and some design features. It is not sufficient to determine whether the intervention groups differed, how precise any estimated differences were, or how broadly the results might apply. Those conclusions require outcome estimates and analytical details not reported in the registry record.
13. Why This Trial Matters Statistically
Statistically, this trial matters because it evaluates an intervention aimed at changing clinician and practice behavior rather than directly assigning a medication to individual patients. Such designs raise questions about the level of randomization, the unit of analysis, and the relationship between provider behavior and patient-level measures. The registry identifies cluster randomization, making the distinction between the assignment unit and the patient-level endpoint especially important.
The endpoint is also broader than a single threshold. It includes patients achieving recommended blood-pressure levels and appropriate responses to uncontrolled blood pressure. That combination can reflect both outcomes and care processes, while requiring careful interpretation of its definition and the way records were reviewed. The registry specifies the baseline measurement approach but does not report its values or later estimates.
The comparison includes audit and feedback in both control and incentive groups. This makes the estimand of interest an added-incentive contrast within a quality-improvement context. The multiple incentive arrangements further create more than one relevant comparison. With 15 outcomes and 4 analyses listed, multiplicity and clear reporting of prespecified comparisons are important; the analysis plan is not reported in the registry record.
Finally, the record’s warning against a simple individual-level re-analysis is a useful reminder that a large number of patient records does not erase clustering. Correct uncertainty depends on how assignment and repeated measurement are represented. Since the sponsor estimates and their precision are not reported in the registry record, the design’s statistical questions can be identified, but its empirical answers cannot be evaluated.
14. Limitations
- The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
- Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
Several trial-specific limitations of interpretation arise from the registry record. The primary outcome is described as an unadjusted baseline-period proportion, but its observed value is not reported. The record does not report the treatment-effect estimates, confidence intervals, p-values, or period-specific values needed to assess either direction or precision. The stated hypothesis therefore cannot be evaluated from the recorded information.
The record identifies cluster randomization and explicitly cautions that a simple individual-level re-analysis would be misleading. However, it does not report the cluster-level assignment details, the number of clusters in each analysis group, or the model used to account for within-cluster dependence. The final number of physicians in the repeated-measures analysis is referred to in a participant flow diagram but is not reported in the registry record. The relationship between enrollment and the analyzed population is consequently unclear from this record.
The baseline endpoint is based on electronic health record review of sampled patients. The record does not report details of missing or incomplete records, measurement reliability, or how patient observations were handled when information needed to classify the endpoint was unavailable. It also does not report a detailed protocol, sample-size justification, interim-analysis plan, or analysis plan. Those elements should not be inferred.
Multiple outcomes and analyses are listed, but prespecification and multiplicity handling are not reported. The record also does not give enough information to assess generalizability beyond the participating physicians and practices. These are limitations of what can be established from the registry record, not proof that the trial itself lacked procedures or analyses that are not recorded there.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Financial Incentives to Translate ALLHAT Into Practice: A Randomized Trial — NCT00302718.
- Registry record last updated 2019-01-30; status last verified 2019-01.
- Results first posted 2014-12-15.
- Data retrieved from ClinicalTrials.gov on 2026-11-09.
- Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
NCT00302718, titled “Financial Incentives to Translate ALLHAT Into Practice: A Randomized Trial,” is a completed, randomized, parallel-group study with no masking. The registry lists hypertension, enrollment of 83, and the VA Office of Research and Development as sponsor. The recorded start is 2007-02 and primary completion is 2011-10.
The trial compared physician-level incentives, practice-level incentives, and combined incentives with a no-incentives control. Control and intervention participants received audit and feedback performance reports. The primary measure concerns the proportion of physicians’ patients meeting blood-pressure thresholds or receiving an appropriate response to uncontrolled blood pressure during the baseline period.
The record lists secondary outcomes and describes a repeated-measures longitudinal analysis, while cautioning against simple individual-level re-analysis because the design is cluster-randomized. Primary outcome estimates and detailed analysis results are not reported in the registry record, so the trial’s observed effects and their precision cannot be summarized.